FRANKENSTEIN MORTGAGES-PSAs-CONTRACTS - Securitization & Mortgages!!

DO NOT DELETE 12/2/2009 3:34 AM
ARTICLES
REWRITING FRANKENSTEIN
CONTRACTS: WORKOUT
PROHIBITIONS IN RESIDENTIAL
MORTGAGE-BACKED SECURITIES
ANNA GELPERN*
ADAM J. LEVITIN†
ABSTRACT
Modification-proof contracts boost commitment and can help

overcome information problems. But when such rigid contracts are
ubiquitous, they can function as social suicide pacts, compelling
enforcement despite significant externalities. At the heart of the current
financial crisis is a contract designed to be hyperrigid: the pooling and
servicing agreement (“PSA”), which governs residential mortgage
securitization. The PSA combines formal, structural, and functional
barriers to its own modification with restrictions on the modification of
underlying mortgage loans. Such layered rigidities fuel foreclosures, with
spillover effects for homeowners, communities, financial institutions,
* Associate Professor of Law, American University Washington College of Law.

† Associate Professor of Law, Georgetown University Law Center. The authors would like to
thank Ian Ayres, Joshua Blank, Susan Block-Lieb, William Bratton, Kevin Davis, Eric Gerding, Mitu
Gulati, John Hunt, Gregory Klass, Sarah Levitin, Ronald Mann, Edward Morrison, David Reiss, Bruce
Spiva, Eric Stein, George Triantis, Elizabeth Warren, Alan White, David Zaring, participants in the
Rutgers-Newark faculty colloquium, the Stanford/Yale Junior Faculty Forum, and the 2009 meeting of
the Canadian Law and Economics Association for their comments and encouragement, and Kelly
Targett for research assistance. Sections of this paper derive from Professor Levitin’s testimony before
the Senate and House Judiciary Committees. The views expressed in this Article are solely those of the
authors.
1075
1076 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 82:1075
financial markets, and the macroeconomy.

This Article situates PSAs in the context of theoretical and policy
debates about contract rigidity, bond contract modification, and
contractual bankruptcy. We propose a typology of contract rigidities,
ranging from formal prohibition on amendment (formal rigidity) to extreme
collective action problems (functional rigidity). We then draw on New Deal
jurisprudence for strategies to overcome each type of rigidity. These
strategies include narrowly tailored legislation that renders the
problematic terms unenforceable on public policy grounds, administrative
restructuring mandates, and special bankruptcy regimes.
The New Deal experience highlights the spillover effects of
widespread contract practices, the limits of voluntary modification, and the
utility of targeted government mandates to rewrite problematic terms. It
also reveals the limits of such mandates. When different kinds of rigidity
combine in a complex web of contracts, a comprehensive mechanism like
bankruptcy may be necessary, if not always sufficient, to break the logjam.
Rewriting PSAs will not resolve today’s financial crisis. Yet voluntary
foreclosure prevention initiatives are unlikely to succeed as long as
contract rigidities persist. The continuing foreclosure epidemic also holds
an important lesson for the future: even where contract rigidity makes
perfect sense for the parties, pervasive rigidities can have catastrophic
consequences for financial stability and for society.
TABLE OF CONTENTS
I. INTRODUCTION: CONTRACT ALCHEMY .................................. 1077
II. THE SPECIES: RMBS ...................................................................... 1080
III. SECURITIZATION’S RIGIDITIES ................................................ 1087
A. FORMAL RIGIDITY 1: PSA LIMITATIONS ON LOAN
MODIFICATION ........................................................................ 1089
B. FORMAL RIGIDITY 2: STATUTORY AND CONTRACTUAL
VOTING THRESHOLDS ............................................................. 1091
C. STRUCTURAL RIGIDITY: BANKRUPTCY REMOTENESS AND
PASSIVE MANAGEMENT .......................................................... 1093
D. FUNCTIONAL RIGIDITY: COORDINATION PROBLEMS FROM
TRANCHING, RESECURITIZATION, AND INSURANCE ............... 1098
E. SUMMARY: RIGID BY DESIGN ................................................. 1102
IV. THE GENUS: RMBS AS IMMUTABLE CONTRACTS ............... 1112
A. FORM, STRUCTURE, AND FUNCTION IN CONTRACT
IMMUTABILITY ........................................................................ 1112
2009] REWRITING FRANKENSTEIN CONTRACTS 1077
B. RIGID CONTRACTS IN BANKRUPTCY THEORY ........................ 1118

V. SPILLOVERS: CONTRACTS AS SUICIDE PACTS...................... 1124
VI. OVERCOMING RIGIDITIES: NEW DEAL TOOLS .................... 1127
A. GOLD CLAUSES AND FORMAL RIGIDITY ................................ 1129
B. UTILITY HOLDING COMPANIES AND STRUCTURAL RIGIDITY. 1135
C. FARM MORTGAGES: A FAILED RESPONSE TO FUNCTIONAL
RIGIDITY .................................................................................. 1141
D. FARM MORTGAGES REVISITED: CHAPTER 12 OF THE
BANKRUPTCY CODE ................................................................ 1148
VII. CONCLUSION: BREAKING THE SPELL ................................... 1149
I. INTRODUCTION: CONTRACT ALCHEMY
Yet you, my creator, detest and spurn me, thy creature, to whom thou art
bound by ties only dissoluble by the annihilation of one of us.1
At the heart of the global financial collapse are two kinds of rigid
contracts. The first is the residential mortgage contract, which becomes
problematic when too many homeowners cannot pay what they owe and
yet cannot modify their debts. The second is the pooling and servicing
agreement (“PSA”), which governs the management of securitized
mortgage loan pools. PSAs are designed to preclude or severely constrain
the modification of both the securitization arrangement and the underlying
mortgages. Both mortgage contract and PSA rigidities can fuel foreclosures
on a large scale, with spillover effects on communities, financial
institutions, financial markets, and the macroeconomy.
To date, the mortgage contract has drawn the bulk of media, policy,
and academic attention.2 We focus instead on the contractual framework
for mortgage securitization and offer an early attempt to integrate the
ongoing crisis into contract theory.
Our case study of mortgage securitization adds a critical dimension to
theories about rigid contracts. Recent contract scholarship has stressed the
welfare-enhancing properties of modification-proof contracts: boosting
1. MARY WOLLSTONECRAFT SHELLEY, FRANKENSTEIN; OR, THE MODERN PROMETHEUS 72

(Susan J. Wolfson ed., Pearson Educ., Inc. 2007) (1818).
2. See, e.g., Helping Families Save Their Homes in Bankruptcy Act of 2009 and the Emergency
Homeownership and Equity Protection Act: Hearing on H.R. 200 and H.R. 225 Before the H. Comm.
on the Judiciary, 111th Cong. (2009); Oren Bar-Gill, The Law, Economics and Psychology of Subprime
Mortgage Contracts, 94 CORNELL L. REV. 1073 (2009); Stan Liebowitz, New Evidence on the
Foreclosure Crisis, WALL ST. J., July 3, 2009, at A13; Eric Posner & Luigi Zingales, The Better,
Cheaper Mortgage Fix: How to Renegotiate All Those Bad Loans at No Cost to the Taxpayer, SLATE,
Mar. 2, 2009, http://www.slate.com/id/2212649/.
commitment between parties and helping them overcome information

asymmetries and agency problems.3 The literature and policy debates on
bond restructuring have struggled with similar arguments for decades.4
Contractual rigidity is also an essential, though implicit, element in the
extensive bankruptcy literature about private contractual ordering of
financial distress.5 Rigidity is necessary to ensure that creditors receive the
contractual bankruptcy regime for which they have bargained.6
We argue that rigidity’s welfare valence is more complex. While
rigidity may be sensible among the contracting parties, when it becomes
widespread in financial contracts, it can produce catastrophic externalities.
Residential mortgage securitization offers an example.
Securitization has been described as financial alchemy, a process that
can change unremarkable financial assets into valuable ones, like lead into
gold.7 Although medieval alchemists contributed much to science and
industry, they never made gold and failed to achieve their grandest promise
of eternal life. Recent experience in the largest securitization market,
residential mortgage-backed securities (“RMBS”), conjures up less
wholesome tales of scientific progress.
In Mary Shelley’s novel, Victor Frankenstein, a Swiss student reared
on alchemy before turning to “real” science, dreamt of animating a new
species.8 He collected parts of corpses from charnel houses to make a
beautiful giant (small parts would take too long).9 But when it came to life,
the being turned hideous and murderous—just human enough to yearn for
society, yet not human enough to join it—and in the end destroyed his
creator and all that he loved.10 Frankenstein’s story, with its themes of
hubris-tainted brilliance, of ordinary substance magically transformed, of
lost control and immutable ties, holds lessons for the immutable contracts
that underlie RMBS and the crisis they have fueled.
Like Frankenstein’s doomed masterpiece, securitization contracts
combine elements of the ordinary (such as high-threshold voting
requirements to change payment terms, common in corporate bonds) with
3. See infra Part IV.A.

4. See infra Part IV.B.
7. See, e.g., Steven L. Schwarcz, The Alchemy of Asset Securitization, 1 STAN. J.L. BUS. & FIN.
133, 134 (1994).
8. See SHELLEY, supra note 1, at 23–36.
9. See id. at 34.
10. See id. at 38, 176.
magical features of financial alchemy (such as bankruptcy remoteness,

tranching, and resecuritization). The result is a layering of rigidities
designed to produce a species of hyperrigid contracts that boost
commitment in good times but function as suicide pacts in bad times.
Our Article begins in Part II with an account of RMBS design. Part III
highlights features of the design that can serve as legal and practical
obstacles to contract modification. Securitization contracts can be rigid in
three ways: formal (contractual prohibition on amendment), structural
(bankruptcy-remote, tax-exempt, and off–balance sheet organization), and
functional (barriers to coordination).
Part IV puts the case study in theoretical context. It situates the PSA
as an exponent of contract rigidity in the literature on bilateral and bond
contract modification. We draw on the writing about asset securitization to
show that—consistent with contract theory—layered rigidity is a key
element of RMBS design. We then link the discussion of contract rigidity
to the bankruptcy theory debate about so-called contractual bankruptcy—
proposals to let firms and creditors choose a private regime for dealing with
business failure. In the RMBS case study, the contracting parties have
devised just such a bankruptcy-remote private regime. Where it works as
designed, this regime makes contract renegotiation very costly for the
parties and, we argue, for society at large. Part V elaborates the far-
reaching effects of contract rigidity.
Part VI surveys techniques for overcoming the three different kinds of
rigidity embodied in the design of RMBS PSAs. After a brief overview of
contract modification proposals in connection with the current financial
crisis, we look back at three examples of New Deal legislation that revised
private contracts, and review the surrounding jurisprudence. These New
Deal laws were enacted, respectively, to take the United States off the gold
standard, to restructure public utilities, and to stop farm foreclosures. Each
sought to overcome a distinct set of contract obstacles that has parallels in
RMBS. The legal literature has yet to reassess the New Deal’s legislative
programs in light of the ongoing crisis. As the country searches for new
paradigms to animate crisis governance, a critical appraisal of specific New
Deal programs is an important place to start.
The New Deal experience highlights the spillover effects of pervasive
contract practices, the limits of voluntary modification initiatives in the
face of collective action problems, and the utility of targeted government
mandates to rewrite problematic terms. It also reveals the limits of such
mandates. When different kinds of rigidity combine in a complex web of
contracts, a comprehensive restructuring mechanism like bankruptcy may

be necessary to break the logjam. Where such contracts are ubiquitous,
even bankruptcy may not be enough.
We deliberately stop short of a New Deal–style legislative proposal to
promote contract modification. Many interesting ideas are on the table at
this writing; no doubt more will emerge as we go to print. None takes a
comprehensive view of the rigidity problem. We argue that the problem is
not susceptible to a silver bullet but rather sits at the heart of the intractable
relationship among contract, bankruptcy, and crisis policy.
Rewriting hyperrigid PSAs will not resolve the financial crisis. Yet
voluntary foreclosure prevention initiatives are unlikely to succeed as long
as contract rigidities persist. The magical features of securitization have
animated a species of contracts designed to bind people, markets, and
governments with “ties only dissoluble by the annihilation of one of us.”11
The continuing foreclosure epidemic holds an important lesson for the
future: even where rigidity makes perfect sense for the contracting parties,
widespread barriers to modification can unleash catastrophic social
consequences. The ex ante benefits of modification-proof contracts must be
weighed against their potential ex post systemic costs. A viable set of tools
to overcome formal, structural, and functional rigidities is essential for
financial stability.
II. THE SPECIES: RMBS
RMBS are a major part of U.S. capital markets. The principal amount
of RMBS outstanding exceeds that of both U.S. corporate bonds and U.S.
Treasury debt.12 As of the end of 2008, there were nearly $6.8 trillion in
outstanding RMBS, accounting for nearly one-quarter of the U.S. bond
market.13 Of the outstanding RMBS, 64 percent were issued by
government-sponsored entities (“GSEs”), the Federal National Mortgage
11. Id. at 72.

12. See Financial Industry Regulatory Agency, The Bond Market, http://apps.finra.org/
investor_information/smart/bonds/401000.asp [hereinafter FINRA] (listing mortgage-related bonds as
$8.9 trillion, corporate bonds as $6.3 trillion, and U.S. Treasury bonds as $5.9 trillion in a bond market
of more than $33.5 trillion) (last visited Oct. 24, 2009). Of the $8.9 trillion in mortgage-related bonds,
over $6.8 trillion are RMBS. See id.; 2 INSIDE MORTGAGE FIN., 2009 MORTGAGE MARKET
STATISTICAL ANNUAL 10 (2009).
13. See supra note 12. The remainder are commercial mortgage-backed securities (“CMBS”).
The true share of mortgage-related securities may be higher because home equity loans and home
equity lines of credit are usually categorized as “asset-backed” rather than “mortgage-backed”
securities. See VINOD KOTHARI, SECURITIZATION: THE FINANCIAL INSTRUMENT OF THE FUTURE 352
(John Wiley & Sons (Asia) Pte Ltd ed. 2006).
Association (“Fannie Mae”) or the Federal Home Loan Mortgage

Corporation (“Freddie Mac”), and carry a guaranty of timely payment of
principal and interest from those entities;14 27 percent were issued through
private conduits;15 and 8 percent are guaranteed by the Government
National Mortgage Association (“Ginnie Mae”) and are backed by the full
faith and credit of the U.S. government.16
RMBS are also central to U.S. housing finance. About 60 percent of
all outstanding residential mortgages by dollar amount are securitized.17
The share of securitized mortgages by number of contracts outstanding is
much higher because the securitization rate is lower for larger “jumbo”
mortgages.18 Over 90 percent of mortgages originated in recent years have
been securitized.19
Residential mortgage securitization transactions are complex and
varied,20 but their core structure is simple.21 A financial institution (the
14. Most, but not all, GSE RMBS carry this guaranty. Some older series of Freddie Mac
participation certificates, for example, only guarantee timely payment of interest and eventual payment
of principal.
15. Private-label RMBS include all securitizations of subprime, scratch-and-dent, and “jumbo”
mortgages (prime mortgages that are larger than the statutory conforming loan limit for the GSEs), and
almost all alt-A securitizations.
16. 12 U.S.C. § 1721(g)(1) (2006) (Ginnie Mae full faith and credit guaranty); 2 INSIDE
MORTGAGE FIN., supra note 12, at 10 (market shares). The GSEs are now in federal conservatorship,
and their obligations carry an “effective guarant[y]” from the federal government, see, e.g., Dawn
Kopecki, Fannie, Freddie Have “Effective” Guarantee, FHFA Says, BLOOMBERG.COM, Oct. 23, 2008,
http://www.bloomberg.com/apps/news?pid=20601087&sid=aO5XSFgElSZA (quoting Federal Housing
Finance Agency Director James Lockhart), but do not enjoy a full faith and credit backing, see 12
U.S.C. § 1719(e) (stating explicitly that GSE debts are not government debts). The difference, if any,
between the “effective guaranty” and “full faith and credit” is uncertain.
17. See BD. OF GOVERNORS OF THE FED. RESERVE SYS., FLOW OF FUNDS ACCOUNTS FOR THE
UNITED STATES: FLOWS AND OUTSTANDINGS, FIRST QUARTER 2009, at 94 tbl.L.218 (2009).
18. See 2 INSIDE MORTGAGE FIN., supra note 12, at 13.
19. Id.
20. For a more detailed description of RMBS transactions, see Adam J. Levitin & Tara Twomey,
Mortgage Servicing 9–68 (Georgetown Univ. Law Ctr., Bus., Econ., & Regulatory Pol’y Working
Paper Series, Paper No. 1324023, 2009) (on file with authors).
21. There are significant distinctions within the RMBS market. The most essential is the
difference in credit risk among Ginnie Mae, GSE, and private-label RMBS. Investors in Ginnie Mae
and GSE RMBS are assuming interest rate risk, not credit risk; investors in private-label RMBS are also
assuming credit risk. This Article is generally focused on private-label RMBS, as Ginnie Mae and GSE
RMBS have more complex regulatory systems and greater flexibility for renegotiation because
nonperforming loans can be purchased out of the securitized pools by the securitization sponsor, thus
removing securitization-related obstacles to loan renegotiation.
Another distinction exists between true securitization and synthetic securitization through the use
of credit derivatives. This Article only addresses true securitizations. For an explanation of synthetic
securitizations, see Ian Bell & Petrina Dawson, Synthetic Securitization: Use of Derivative Technology
for Credit Transfer, 12 DUKE J. COMP. & INT’L L. 541 (2002).
“sponsor”) owns a pool of mortgage loans, which it either made itself

(“originated”) or bought. Rather than hold these mortgage loans and the
credit risk on its own books, it sells them to a shell entity, a special purpose
vehicle (“SPV”) that is typically structured as a trust. The trust raises the
funds to pay for the loans by issuing securities, which are much like bonds
whose payments are secured by the loans in the trust.22 These are known as
RMBS. Figure 1 illustrates a simplified prototypical private-label RMBS
transaction.
FIGURE 1. Prototypical Private-Label RMBS Transaction
Frequently, the sale of the loans from the sponsor to the SPV involves
an intermediate sale from the sponsor to a wholly owned subsidiary (the
“depositor”) in order to segregate the assets from the sponsor’s other assets
22. See KOTHARI, supra note 13, at 10–12; STEVEN L. SCHWARCZ, BRUCE A. MARKELL & LISSA
LAMKIN BROOME, SECURITIZATION, STRUCTURED FINANCE AND CAPITAL MARKETS 2–3 (2004).
and liabilities for bankruptcy remoteness and accounting consolidation

purposes.23 The depositor then resells the loans to the SPV in exchange for
the RMBS, which the depositor sells to an underwriting affiliate of the
sponsor. The underwriting affiliate pays for the RMBS by reselling them to
investors. These steps all take place in an integrated transaction. Figure 2
illustrates the transaction from figure 1, with these additional steps added.
FIGURE 2. Prototypical Private-Label RMBS Transaction: Expanded
23. See infra Part III.C.

Institutions have many different and overlapping reasons to

securitize.24 Some are well placed to make (originate) loans but do not
want to hold long-term credit risk on their books.25 By securitizing, they
seek to transfer the credit risk to the investors in the securities issued by the
SPV.26 Such institutions make money off up-front fees rather than interest
payment streams. Securitization turns delayed payment streams, like
periodic loan payments, into up-front cash. Securitization thus increases
liquidity, which enables more lending.27
Others seek to reduce their overall cost of funds.28 Raising money
through securitization can be cheaper than taking out loans or issuing
securities directly because the borrowing cost is based on the quality of the
transferred assets, not the overall risk profile of their seller.29
Securitization can further reduce borrowing costs through financial
engineering. Techniques such as the division of the SPV’s securities into
senior and subordinate “tranches” expand the potential investor base.30
They allow the SPV to target new investors with tailored payment
structures and credit enhancements.31 In particular, they permit the issuance
of some securities at a higher credit rating than the overall quality of the
assets in the SPV.32 Such senior securities can be sold to institutional
investors that may only buy investment-grade paper.33 Adding potential
investors boosts overall demand and lowers the cost of financing.34 See
24. See, e.g., COMPTROLLER OF CURRENCY, U.S. DEP’T OF TREASURY, ASSET SECURITIZATION:
COMPTROLLER’S HANDBOOK 2, 4–5 (1997) [hereinafter OCC HANDBOOK]. For a detailed discussion of
the advantages of securitization to issuers, see KOTHARI, supra note 13, at 97–102. For a discussion of
why banks securitize, see Fed. Reserve Bd., Trading and Capital-Markets Activities Manual § 3020.1,
at 1–2 (Jan. 2009), available at http://federalreserve.gov/boarddocs/supmanual/trading/200901/
0901trading.pdf.
25. See KOTHARI, supra note 13, at 100–01.
26. See id. at 100–01, 104–05.
27. Steven L. Schwarcz, Securitization Post-Enron, 25 CARDOZO L. REV. 1539, 1560 (2004).
28. KOTHARI, supra note 13, at 97, 100.
29. See Schwarcz, supra note 7, at 136.
30. See KOTHARI, supra note 13, at 97. The typical subprime residential private-label RMBS
deal has fifteen tranches. See Ingo Fender & Peter Hördahl, Estimating Valuation Losses on Subprime
MBS with the ABX HE Index—Some Potential Pitfalls, BIS Q. REV., June 2008, at 6, 7 n.7, available at
http://www.bis.org/publ/qtrpdf/r_qt0806.pdf.
31. See Claire A. Hill, Securitization: A Low-Cost Sweetener for Lemons, 74 WASH. U. L.Q.
1061, 1075 (1996).
32. See id. at 1073.
33. Id. at 1071.
34. The originator need not, and does not normally, pass on all its cost savings from
securitization to its borrowers. Thus, if the securitized assets are mortgage loans with an annual yield of
8 percent, but the average coupon promised to the SPV investors is 6 percent, the originator can capture
some of the 2 percent spread.
Figure 3 for an illustration of senior/subordinate tranching. Typically,

investment-grade tranches make up the bulk (95 percent or more) of a
private-label RMBS issuance.35
FIGURE 3. Tranching of RMBS
Securitization can also bring accounting and tax benefits. When the
originator sells assets to an SPV, it both moves them off its balance sheet
and recognizes future revenue up front (gain-on-sale accounting).36 Moving
risky and poorly performing assets off the books makes the books look
better and, for financial institutions, lowers regulatory reserve
requirements.37 Moreover, the sale of a potential future revenue stream
35. See Adam B. Ashcraft & Til Schuermann, Fed. Reserve Bank of N.Y., Staff Report No. 318,
Understanding the Securitization of Subprime Mortgage Credit 30 (Mar. 2008), available at
http://www.newyorkfed.org/research/staff_reports/sr318.pdf.
36. See KOTHARI, supra note 13, at 100. See also id. at 102 (noting that improved accounting
profits “might well top the list” of the benefits of securitization).
37. See Hill, supra note 31, app. D, at 1123–25 (mentioning the “capital adequacy ratio”
regulatory requirement, which requires institutions to maintain a certain amount of capital relative to
their risk-weighted assets).
(like loan payments) for up-front cash may produce an immediate revenue
boost on the originator’s books, which can be appealing in many respects,
including when executive compensation is pegged to short-term results.
And some securitization structures, especially for RMBS, are designed to
avoid federal entity-level taxation for the SPV, so that only the investors’
income is subject to federal taxation.38 Avoiding double taxation makes
RMBS and similar instruments more attractive to investors than ordinary
corporate debt securities and further lowers the cost of funds for
originators.
In sum, securitization can be used to achieve a broad range of goals,
including the ordinary business objectives summarized above, as well as
others, ranging from political risk management39 to fraud.40 Securitization
is a financial engineering tool with many uses. But with the possible
exception of fraud, achieving an originator’s objectives requires the
transfer of credit risk to someone else.41
The problem is that most investors in RMBS or any other asset-backed
securities do not want a given homeowner’s credit risk any more than the
originator does. And they certainly do not want the originator’s credit
risk.42 Securitization has responded with solutions in the form of insurance
and immutability. We discuss insurance below; we address immutability in
Part III.
Insurance can come in several nonexclusive forms, commonly referred
to as “credit enhancements.”43 First, insurance can take the form of a
guarantee from the originator. This wholly or partially defeats the point of
38. See OCC HANDBOOK, supra note 24, at 18–19; SCHWARCZ ET AL., supra note 22, at 113–17
(discussing two securitization vehicles recognized by the federal tax code that specifically avoid entity-
level tax); infra note 85.
39. See generally Claire A. Hill, Latin American Securitization: The Case of the Disappearing
Political Risk, 38 VA. J. INT’L L. 293 (1998) (analyzing the use of cross-border securitization to reduce
investors’ political risk exposure).
40. Enron’s structured vehicles are the most famous example.
41. This risk transfer is part of the “true sale” concept. See, e.g., SECURITIZATION OF FINANCIAL
ASSETS § 2.02, at 2-14 (Jason H. P. Kravitt & Mayer Brown eds., 2d ed. Supp. 2008) (broadly defining
a “true sale” as “a transfer of financial assets that, for the purpose of specific laws, accounting
principles or regulatory concerns, constitutes a sale of such assets as distinguished from a financing of
the seller thereof secured by such assets”). See also SCHWARCZ ET AL., supra note 22, at 6–8, 145;
Kenneth C. Kettering, Securitization and Its Discontents: The Dynamics of Financial Product
Development, 29 CARDOZO L. REV. 1553, 1585–1632 (2008) (arguing that securitization is inherently a
fraudulent transfer).
42. Investors are expecting to carry interest rate and other market risk. OCC HANDBOOK, supra
note 24, at 13, 19–20. See also SCHWARCZ ET AL., supra note 22, at 6–8, 145.
43. OCC HANDBOOK, supra note 24, at 11, 19–23.
the transfer for the originator and leaves the investors with the originator’s
credit risk.
Second, insurance can be embedded in the structure of the
securitization vehicle. For example, some securitizations are designed so
that the SPV receives cashflows beyond what is needed to pay the bond
coupons (overcollateralization). Such excess cash is escrowed to cover
potential shortfalls of incoming funds. This cash may or may not be enough
to cover the risk to investors. In either case, the originator may not want to
bear the cost of overcollateralizing the SPV, which involves having funds
sit in escrow, yielding a low return.
Third, the senior/subordinate tranching discussed earlier can be
understood as a form of insurance that some investors provide to the others:
if the SPV does not generate sufficient cashflows to pay all creditors, the
subordinate creditors agree to absorb losses from reduced cashflows first,
up to the amount due to them.44
Finally, insurance can come from a third party, such as an insurance
company that guarantees some or all payments to the investors if the SPV
fails to pay.
Insurance alone does not eliminate the problem of credit risk; it simply
shifts the risk to the insurer. Insurance can be expensive or partial, but even
where it is cheap and comprehensive, it is only as good as the credit of the
insurer. Part III discusses another approach to managing credit risk—
immutability—and other less drastic barriers to payment modification.
III. SECURITIZATION’S RIGIDITIES
Contracts that cannot be modified are more likely to perform

according to their original terms. This is the idea behind the quest for
immutability as a method of managing risks from securitization. We
discuss the theoretical underpinnings of immutability in the next part.
Below we elaborate on the role of immutability in securitization.
Originators opt to securitize, rather than simply sell, financial assets
where “the ‘loss’ on each contract cannot be known with certainty” because
buyers “will discount what they are willing to pay [for any given
asset] . . . . If, however, [the originator] can assure a return to a buyer or
investor, the deal may be worthwhile.”45 Legal tools to assure investors
44. Id. at 11. We discuss tranching in more detail in Part III.D infra.
45. SCHWARCZ ET AL., supra note 22, at 9.
such a specific return seek to lock in a predictable pattern of behavior on

the part of all those involved in securitization whose actions might affect
cashflows, including actors that are not party to the securitization
contracts—for example, homeowners, third-party creditors, and
government officials. Contract rigidity in our account is a multilayered
private arrangement to maximize such lock-in.
Securitization contracts have many attributes designed to make
modification costly and difficult. This part explains such attributes, and
classifies them into a typology of rigidities: formal, structural, and
functional. We consider the barriers to amending contracts governing
mortgage pool securitization separately from the barriers to amending the
underlying loan contracts. Although mortgage contracts are quite rigid in
their own right,46 we focus on pool-level rigidities and limit the discussion
of loan-level rigidities to those arising from securitization arrangements.
Barriers to amendment in securitization contracts generally respond to
agency concerns. RMBS have myriads of investors in pools of thousands of
mortgage loans. Transaction costs make it impractical for the investors to
manage the underlying loan portfolio. Their debtor—the trust that owns the
loans—is an inanimate shell that does not make much of a manager. The
solution is for the investors to hire an agent, called a servicer, to administer
the loan pool: to send out bills, allocate payments, dun delinquent
homeowners, and foreclose on homes where the loan is in default.47
Delegating management to the servicer in turn creates agency risks for
investors, including the risk that the servicer will renegotiate the underlying
loans, reducing payments to the pool, for example, in exchange for a side
payment.
Investors address agency risk contractually through the PSA between
the servicer and the trust.48 PSAs typically direct servicers to manage the
46. A large literature attests to this rigidity. Obstacles to mortgage modification include
prepayment penalties, see, e.g., Frank S. Alexander, Mortgage Prepayment: The Trial of Common
Sense, 72 CORNELL L. REV. 288 (1987), statutory restrictions on mortgage modification in bankruptcy,
see Adam J. Levitin, Resolving the Foreclosure Crisis: Modification of Mortgages in Bankruptcy, 2009
WIS. L. REV. 565, 579–82, and the existence of multiple liens on the same property, which precludes
senior mortgagees from making concessions that would benefit junior mortgagees before improving the
debtor’s payment capacity, see id.
47. OCC HANDBOOK, supra note 24, at 10. The servicer is often, but not always, a corporate
affiliate of the originator. Many large servicers are subsidiaries of bank holding companies.
48. Three kinds of agreements form the core of a securitization transaction: a pooling agreement,
in which the SPV purchases a pool of assets from the originator or an intermediary; a servicing
agreement between the servicer and the SPV that sets forth the duties and compensation of the servicer;
and an indenture, which sets forth the rights of the investors in the SPV’s securities and the duties of the
trustee that oversees the securities and the SPV. Typically, these three agreements are combined into a
loans owned by the SPV as they would their own loans,49 subject to
specific limits on discretion to mitigate loan losses. Thus, where a lender,
maximizing recovery for its own account, might renegotiate a distressed
mortgage loan, a PSA may preclude the servicer from doing the same by
methods ranging from formal prohibition to contracting practices that
produce creditor coordination problems. To lock in such limits, the PSA
further constrains its own modification using similar methods. Loan- and
pool-level constraints combine in a layered structure that is designed to be
unusually rigid.
Below we detail the three kinds of rigidity layered in securitization
arrangements: explicit restrictions on amendment, rigidities that stem from
the organizational form of the SPV, and those that stem from creditor
collective action problems.
A. FORMAL RIGIDITY 1: PSA LIMITATIONS ON LOAN MODIFICATION
PSA terms often explicitly limit modification of the underlying

mortgage loans. Loan modification limitations are designed to restrict the
discretion of the servicer so as to mitigate agency risk in securitizations.50
single document (the PSA). See id. at 13 (noting that the PSA generally governs the structure of the
securitization transaction). See also DB Structured Prods., Inc. v. Am. Home Mortgage Holdings, Inc.
(In re Am. Home Mortgage Holdings, Inc.), 402 B.R. 87, 100 ¶ 40, 103 ¶ 49 (Bankr. D. Del. 2009)
(holding that servicing rights were severable from a mortgage loan purchase agreement and therefore
were assignable by the bankrupt servicer rather than forfeited).
49. See, e.g., Asset Backed Funding Corp., Option One Mortgage Corp. & Wells Fargo Bank,
Pooling and Servicing Agreement (Form 8-K), at EX-4 § 3.01 (Nov. 25, 2005), available at
http://www.secinfo.com/dRSm6.z251.d.htm#6f6m [hereinafter ABFC PSA] (“The Servicer, as
independent contract servicer, shall service and administer the Mortgage Loans in accordance with this
Agreement and the normal and usual standards of practice of prudent mortgage servicers servicing
similar mortgage loans and, to the extent consistent with such terms, in the same manner in which it
services and administers similar mortgage loans for its own portfolio, and shall have full power and
authority, acting alone, to do or cause to be done any and all things in connection with such servicing
and administration which the Servicer may deem necessary or desirable and consistent with the terms of
this Agreement (the ‘Servicing Standard’).”); Goldman Sachs Mortgage Co. & Bank One, N.A.,
Seller’s Purchase, Warranties and Servicing Agreement (Form 8-K), at EX-10.1.3 § 4.01 (Mar. 8,
2002), available at http://www.sec.gov/Archives/edgar/data/807641/000095017202000467/s575865.txt
(“The Servicer shall service and administer the Mortgage Loans through the exercise of the same care
that it customarily employs for its own account.”).
50. Servicers are required to advance payments of principal and interest on nonperforming loans
held by the SPV. The advances are reimbursable, but not with interest. One agency concern is that
servicers might be tempted to modify nonperforming loans in order to avoid the obligation of making
advances, even if the modified loan has a lower net present value than would result if it proceeded to
foreclose on the nonperforming loan. Of course, the opposite situation—where modification would
yield a greater return than foreclosure—could well hold.
Sometimes modification is forbidden outright;51 sometimes renegotiation is

permitted only under limited circumstances;52 sometimes only certain types
of renegotiation are permitted;53 and sometimes third-party consent is
required to renegotiate loans beyond a specified cap (typically 5 percent of
the pool).54 PSAs occasionally limit the number of renegotiations by loan
or by year.55 Additionally, servicers may be required to purchase any loans
they renegotiate at the face value outstanding or at a premium.56 Imposing
51. See, e.g., Federal National Mortgage Ass’n, Single-Family Master Trust Agreement for
Guaranteed Mortgage Pass-Through Certificates § 5.3(4) (June 1, 2007), available at
http://www.fanniemae.com/mbs/pdf/singlefamilytrustagreement_June2007.pdf (“For so long as a
Mortgage Loan remains in a Pool, the Mortgage Loan may not be modified if the modification has the
effect of changing the principal balance (other than as a result of a payment actually received from or
on behalf of the Borrower), changing the Mortgage Interest Rate (other than in accordance with any
adjustable rate provisions stated in the Mortgage Documents), or delaying the time of payment beyond
the last scheduled payment date of that Mortgage Loan . . . .”). Note that this is not a private-label
securitization and that Fannie Mae can purchase defaulted loans out of its securitized pools and modify
them.
52. See, e.g., ABFC PSA, supra note 49, § 3.03 (“In the event that any payment due under any
Mortgage Loan is not paid when the same becomes due and payable, or in the event the Mortgagor fails
to perform any other covenant or obligation under the Mortgage Loan and such failure continues
beyond any applicable grace period, the Servicer shall take such action as it shall deem to be in the best
interest of the Certificateholders. With respect to any defaulted Mortgage Loan, the Servicer shall have
the right to review the status of the related forbearance plan and, subject to the second paragraph of
Section 3.01, may modify such forbearance plan; including extending the Mortgage Loan repayment
date for a period of one year or reducing the Mortgage Interest Rate up to 50 basis points.”).
53. Morgan Stanley Capital I Inc./Trust 2006-HE1, Pooling and Servicing Agreement (Form 8-
K), at EX-4 § 3.01(c) (May 11, 2006) (“Notwithstanding anything in this Agreement to the
contrary . . . the Servicer shall not (i) permit any modification with respect to any Mortgage Loan that
would change the Mortgage Rate, reduce or increase the principal balance (except for reductions
resulting from actual payments of principal) or change the final maturity date on such Mortgage
Loan . . . .”); Morgan Stanley Capital I Inc./Trust 2006-NC2, Pooling and Servicing Agreement (Form
8-K), at EX-4 § 3.01(c) (May 11, 2006) (same). We thank John Hunt for bringing our attention to these
particularly restrictive PSAs.
54. See, e.g., ABFC PSA, supra note 49, § 3.01 (“The [net interest margin security] Insurer’s
prior written consent shall be required for any modification, waiver or amendment if the aggregate
number of outstanding Mortgage Loans which have been modified, waived or amended exceeds 5% of
the number of Mortgage Loans as of the Cut-off Date.”).
55. CREDIT SUISSE, THE DAY AFTER TOMORROW: PAYMENT SHOCKS AND LOAN
MODIFICATIONS 7 (2007).
56. See, e.g., CWALT Inc., Prospectus Supplement, Alternative Loan Trust 2005-1cb: Mortgage
Pass-Through Certificates (Form 4249(b)(5)), at S-60 (Feb. 1, 2005), available at http://www.sec.gov/
Archives/edgar/data/1269518/000095012905000801/v04622b5e424b5.txt (“The master servicer may
modify any mortgage loan, provided that the master servicer purchases the mortgage loan from the trust
fund immediately following the modification. A mortgage loan may not be modified unless the
modification includes a change in the interest rate on the related mortgage loan to approximately a
prevailing market rate. Any purchase of a mortgage loan subject to a modification will be for a price
equal to 100% of the Stated Principal Balance of that mortgage loan, plus accrued and unpaid interest
on the mortgage loan up to the next Due Date at the applicable net mortgage rate, net of any
unreimbursed advances of principal and interest on the mortgage loan made by the master servicer.”);
the full cost of modification on the servicer makes them reluctant to

modify; it is meant to serve as a barrier to renegotiation.
The incidence of such modification restrictions is the subject of
ongoing empirical investigation. One small early sampling by Credit Suisse
found that about one-third of all private-label (non-government-sponsored)
securitizations had some restriction on modification.57 Another study by
Bear Stearns found that 40 percent of the private-label securitizations in a
widely followed index had a 5 percent limit on modifications (absent
approval by a ratings agency) and that 10 percent of the deals permitted no
modification whatsoever.58 A more extensive academic study found an
absolute bar to modification in nearly 10 percent of subprime RMBS
transactions in 2006, and a range of other restrictions in most other deals.59
In the remainder of this part, we argue that the content and incidence
of explicit contractual restrictions on modification of the sort discussed so
far are, at best, a partial and occasionally misleading indication of the
actual modification propensity of any given loan or securitization
arrangement.
B. FORMAL RIGIDITY 2: STATUTORY AND CONTRACTUAL VOTING

THRESHOLDS
The Trust Indenture Act of 193960 (“TIA”) directly impedes

modification of RMBS payment terms, which may in turn deter
modification of the underlying assets. Like most U.S. corporate bonds,
RMBS are subject to the TIA, which requires each investor’s consent to
modify its right to receive principal and interest payments according to the
terms of its security.61 In effect, modifying the economic terms of RMBS
Complaint ¶ 34, Greenwich Fin. Servs. Distressed Mortgage Fund, LLC v. Countrywide Fin. Corp., No.
650474-2008E (N.Y. Civ. Ct. Dec. 1, 2008), available at http://www.businessweek.com/pdfs/2008/
1201_complaint.pdf (quoting CWALT PSA § 3.11(b)) [hereinafter Greenwich Complaint]
(“Countrywide may agree to a modification of any Mortgage Loan (the ‘Modified Mortgage Loan’)
if . . . Countrywide purchases the Modified Mortgage Loan from the Trust Fund . . . .”).
57. See CREDIT SUISSE, supra note 55, at 7.
58. See Vikas Bajaj, For Some Subprime Borrowers, Few Good Choices, N.Y. TIMES, Mar. 22,
2007, at C1.
59. See John P. Hunt, What Do Subprime Securitization Contracts Actually Say About Loan
Modification? Preliminary Results and Implications 2–3, 6–10 (Mar. 25, 2009) (unpublished
manuscript, available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1369286 (surveying the
614 subprime home mortgage securitization deals from 2006).
60. 15 U.S.C. § 77aaa–77bbbb (2006).
61. Id. § 77ppp(b) (“Notwithstanding any other provision of the indenture . . . , the right of any
holder of any indenture security to receive payment of the principal of and interest on such indenture
security, on or after the respective due dates expressed in such indenture security, or to institute suit for
requires the consent of 100 percent of their holders. Unanimity is hard to

achieve because securities are held by a great multitude of creditors
worldwide. Simply failing to contact one can be fatal.
In the first instance, the TIA works as a constraint on amending
RMBS to reduce payments to the creditors, either by altering credit
enhancement arrangements or changing cashflows to reflect a restructuring
of the underlying assets. Indirectly, this constraint on amending RMBS
may in turn limit the servicer’s capacity to renegotiate the underlying
mortgage loans. Most loan modifications—reductions in principal or
interest rates, or extension of maturities—could reduce payments to the
SPV and thereby impair payments to RMBS holders to the point of pushing
the vehicle itself into default.62 The TIA would bar RMBS amendments to
permit such impairment and avoid default, even if avoiding default brought
higher recovery values.
The extent to which TIA strictures actually function as a barrier to
amendment is uncertain. It is not generally cited among the obstacles to
underlying loan modification. There is no case law directly on point. On its
face, the TIA only requires that an investor’s right to payment not be
impaired; it does not govern management of the SPV’s assets any more
than it governs management of corporate assets for corporate bonds.
Protection of the right to payment is not the same as prescribing a means of
generating income to pay the debt.
Yet it is conceivable that a court might treat assets of an SPV—a
passive synthetic creature whose only purpose is to funnel highly specified
cashflows to investors—differently from ordinary corporate assets.63
Uncertainty surrounding the application of the TIA in securitization may in
turn act as a constraint on both mortgage and PSA modification, as risk-
averse servicers will hesitate to test the limits of the TIA and face potential
the enforcement of any such payment on or after such respective dates, shall not be impaired or affected
without the consent of such holder . . . .”).
62. The TIA does not require any particular threshold of consent for other modifications to the
indenture, but most RMBS supererogate and require either a simple majority vote or a two-thirds
majority vote for modifications not affecting cashflow. The TIA does, however, require simple majority
consent for waivers of past defaults. Id. § 77ppp(a)(1).
63. The servicer might argue that the real alternative to modification is not payment on the
original schedule, but unpredictable recovery from foreclosures. Investors might respond that
foreclosure need not compromise the SPV’s—and hence the investors’—right to receive the original
principal and interest on schedule; it is just a means of collecting some of the payments through a sale
of collateral. The argument’s permutations are numerous, and the outcome uncertain, reflecting the
variety of economic and legal arrangements in the market.
litigation.64
Even if the TIA unanimity requirement did not apply, the PSAs
themselves often impose supermajority requirements for PSA amendment.
Most commonly, two-thirds of each affected tranche of RMBS holders
must consent to a modification of the PSA.65 While less daunting than 100
percent, requiring the consent of two-thirds of each affected tranche gives
creditors significant capacity to hold up modification. Such requirements
fuel collective action problems, about which we elaborate below in Part
III.D.66
C. STRUCTURAL RIGIDITY: BANKRUPTCY REMOTENESS AND PASSIVE

MANAGEMENT
In corporate bonds, bankruptcy overcomes creditor coordination

problems.67 If the debtor corporation is unable to renegotiate its bond debt
64. See, e.g., Citibank N.A. v. MBIA Assurance S.A., [2006] EWHC (Ch) 3215, aff’d, [2007]
EWCA (Civ) 11. This recent British case illustrates the deep uncertainty that exists regarding legal
duties in securitization structures, as well as the tranche warfare that workouts can create. In this case, a
Dutch SPV held some of the debt of Eurotunnel (owner of the English Channel tunnel or the
“Chunnel”), against which it had issued bonds. Id. [1]–[3]. Eurotunnel had to restructure its debt
obligations, and the restructuring provided for some of the Eurotunnel debt held by the SPV to be
exchanged for Eurotunnel notes redeemable as Eurotunnel shares plus cash, but it included an option to
exchange the debt for cash alone at 61.9 percent of par value. Citibank, [2007] EWCA (Civ) 11, [6].
The SPV’s indenture provided that the trustee and the credit-enhancing insurer, MBIA, had to approve
such deals. Citibank, [2006] EWHC (Ch) 3215, [7]. MBIA instructed the trustee, Citibank, to exercise
the all-cash option. Id. [18]. The holder of a junior tranche of SPV bonds objected to this deal. Id. [17].
It seems that if the all-cash option were exercised, the junior trancheholder would have been out of the
money because the cash was at 61.9 percent of par value, whereas with the combination of notes
redeemable as shares and cash, the junior trancheholder would have received some value. See id. The
issue before the court was whether MBIA could direct Citibank, as trustee, to exercise the option and, if
not, whose interests Citibank had to consider in evaluating whether to exercise the option. The court
ruled that MBIA had the right under the indenture to direct Citibank’s action; but the case shows that
there is sufficient legal uncertainty about rights and duties in securitization structures, including the
duties of the trustee (like a servicer) to act in the interests of the trust as a whole, given that the junior
trancheholder was willing to litigate in the U.K.’s loser-pays-all-attorney’s-fees litigation system.
65. See, e.g., ABFC PSA, supra note 49, § 11.01 (“[T]his agreement may be amended from time
to time . . . provided, however, that no such amendment or waiver shall (x) reduce in any manner the
amount of, or delay the timing of, payments on the Certificates which are required to be made on any
Certificate without the consent of the Holder of such Certificate, (y) adversely affect in any material
respect the interests of the Holders of any Class of Certificates or the Swap Provider in a manner other
than as described in clause (x) above, without the consent of the Holders of Certificates of such Class
evidencing at least 66 2/3% of the Voting Rights evidenced by such Class . . . .”).
66. Some PSAs permit the trustee or servicer to execute, without RMBS holder consent,
amendments to the PSA necessary to preserve real estate mortgage investment conduits’ pass-through
tax status.
67. THOMAS H. JACKSON, THE LOGIC AND LIMITS OF BANKRUPTCY LAW 7–19 (1986).
consensually, it can file for bankruptcy and force a debt restructuring.68

This possibility creates pressure for a consensual deal. Bankruptcy,
however, is either formally unavailable or practically inaccessible for
RMBS.
A typical securitization SPV is structured so that it cannot file for
bankruptcy.69 It is also shielded from being dragged into the bankruptcy of
the originator.70 Such “bankruptcy remoteness” is achieved using a
combination of organizational form and contract.71 For example, most
SPVs that issue RMBS are organized as trusts because trusts, excluding
“business trusts,” are not “persons” eligible to file for bankruptcy under the
Bankruptcy Code.72 The Bankruptcy Code does not define “business trust,”
and case law is unsettled on the definition,73 so most SPVs take further
steps to achieve bankruptcy remoteness by prohibiting the trust from filing
a voluntary bankruptcy petition, making all parties to the securitization
transaction covenant not to file an involuntary petition against the SPV,74
and limiting the SPV’s activities to avoid third-party creditors who could
file an involuntary petition against the SPV.75 Organizational forms that
qualify as “persons” for bankruptcy purposes may require additional
contractual commitments, such as high-threshold voting requirements to
file a voluntary petition.76
68. See id.

69. See KOTHARI, supra note 13, at 11–12 (describing “bankruptcy remoteness”).
70. See id.
71. See SCHWARCZ ET AL., supra note 22, at 54. Outright advance waivers of bankruptcy rights
are historically disfavored by U.S. courts. See, e.g., In re Weitzen, 3 F. Supp. 698, 698 (S.D.N.Y. 1933)
(“The agreement to waive the benefit of bankruptcy is unenforceable.”).
72. Only “persons” are eligible to file for bankruptcy. See 11 U.S.C. § 109(a) (2006). The
Bankruptcy Code defines “person” to include an “individual, partnership, and corporation,” but it does
not include a “governmental unit.” Id. § 101(41). The definition of “corporation,” however, includes a
“business trust.” Id. § 101(9)(A)(v).
73. See, e.g., Shawmut Bank Conn. v. First Fid. Bank (In re Secured Equip. Trust of E. Air
Lines, Inc.), 38 F.3d 86, 88–91 (2d Cir. 1994); In re Eagle Trust, No. 98-2531, 1998 U.S. Dist. LEXIS
14488, at *12 (E.D. Pa. Sept. 16, 1998) (stating that “[t]he various courts that have addressed the issue
have applied different factors to determine the existence of a business trust”).
74. See, e.g., ABFC PSA, supra note 49, § 3.28(b) (“Each party to this Agreement agrees that it
will not file an involuntary bankruptcy petition against the Trustee or the Trust Fund or initiate any
other form of insolvency proceeding until after the Certificates have been paid.”).
75. See, e.g., id. § 3.28(a) (stating that “the Trust is not authorized and has no power” to “borrow
money or issue debt,” to “merge with another entity, reorganize, liquidate or sell assets,” or to “engage
in any business or activities”).
76. See Stephen H. Case, I Thought I Put That Where You Couldn’t Reach It: Bankruptcy-
Remote Entities, Special-Purpose Vehicles and Other Securitization Issues, in DEALING WITH SECURED
CLAIMS AND STRUCTURED FINANCIAL PRODUCTS IN BANKRUPTCY CASES 66 (2004) (PLI Commercial
Law and Practice, Course Handbook Series 2002).
To make bankruptcy remoteness meaningful, the vehicle must be

protected from the misfortunes both of the originator who sold the
mortgage loans and those of the original debtors. The first of these
objectives is achieved with a “true sale”—ensuring that the originator does
not retain a residual interest in the mortgage loans, so that such interest
does not become an asset of the originator’s bankruptcy estate.77 The
segregation of the SPV’s assets from the originator’s is a major component
of the value of securitization78 and is typically memorialized in two opinion
letters—a “true sale” opinion and a “non-consolidation” opinion from the
law firm representing the securitization transaction’s sponsor.79 The second
objective—insulating investors from the underlying mortgage troubles—is
advanced by the restrictions on loan modification discussed in the
preceding section, and in part through the Bankruptcy Code’s prohibition
on the modification of single-family principal residence mortgages.80
77. KOTHARI, supra note 13, at 11. See also id. at 207–08 (discussing the complexity of
structuring the transfer of receivables).
78. See Case, supra note 76, at 76 (“Because a securitization transaction is based on the premise
that the assets being financed have been isolated from the risks that the transferor/originator will default
on debt or enter bankruptcy, the legal conclusion that the assets after transfer to the [SPV] have ceased
to be the property of the originator and have become the property of the [SPV] are critical for
bankruptcy purposes. . . . Characterization of the transaction as a ‘true sale’ achieves these goals . . . .”);
Kenneth M. Ayotte & Stav Gaon, Asset-Backed Securities: Costs and Benefits of “Bankruptcy
Remoteness” 3 (May 14, 2006) (unpublished manuscript), available at http://www.newyorkfed.org/
research/conference/2006/cffi/Ayotte.pdf (explaining that unlike a debtor’s collateral, which can be
seized “on demand” by credtors, “assets that were transferred in a ‘true sale’ to the SPV are not
considered part of the debtor’s bankruptcy estate, but instead continue to be used for the benefit of the
SPV investors”).
79. KOTHARI, supra note 13, at 208. The concern is that the securitization could be declared a
fraudulent conveyance by the originator and the securitized assets could be consolidated with the
originator’s. In that event, the securitization would be treated as an unperfected secured financing, with
little or no prospect of recovery for the investors. See First Interim Report of Neal Batson, Court-
Appointed Examiner at 43, In re Enron Corp., 2009 Bankr. LEXIS 2094 (Bankr. S.D.N.Y. Aug. 4,
2009) (No. 01-16034 (AJG)), available at http://www.enron.com/media/1st_Examiners_Report.pdf
(“One of the more important considerations in this analysis that is common to many of the Selected
Transactions is the existence of one or more instruments . . . that had the economic effect of transferring
both the obligation to repay the financing and the economic risks and rewards of the ‘sold’ asset back to
Enron.”). See also Case, supra note 76, at 81–83 (noting that fraudulent transfers are avoidable in
bankruptcy proceedings and mentioning ways to ensure that a securitization is not treated as a
fraudulent transfer). Similarly, where the originator is a U.S. bank not subject to the Bankruptcy Code,
securitization requires measures to prevent the assets from being dragged into bank resolution
proceedings by the Federal Deposit Insurance Corporation. See OCC HANDBOOK, supra note 24, at 19–
20 (discussing the structure of various levels of risk in securitization transactions); SCHWARCZ ET AL.,
supra note 22, at 151–54.
80. See 11 U.S.C. § 1322(b)(2) (2006). See generally Levitin, supra note 46 (discussing the
policy assumptions behind the home loan modification prohibition and testing these policy assumptions
empirically).
By contracting out of bankruptcy, the securitization SPV creates

structural, pool-level rigidity absent in ordinary corporate bonds. If the
underlying assets do not generate enough cash to meet the SPV’s liabilities,
the dominant option is a sequence of defaults on the creditor hierarchy.
Voluntary renegotiation of amounts due to creditors is made difficult by the
high-threshold voting requirements described in the previous section81 and
by conflicts of interest between senior and subordinate creditors described
in the next section.82 Bankruptcy-remote organization rules out involuntary
changes.83
The SPV’s bankruptcy-remote organization dovetails with another
structural feature of the pooling arrangement that contributes to loan-level
rigidity: passive management. The SPV’s capacity to avoid entity-level
taxation and to remain off the originator’s books usually hinges on being
passively managed. An actively run business is no mere pass-through.
Similarly, a servicer that really treats the SPV’s assets as its own risks
having regulators do the same.
Most PSAs restrict mortgage renegotiation to loans that are in default
or where default is imminent or reasonably foreseeable84 in order to protect
the SPV’s pass-through tax and off–balance sheet accounting status.
Allowing modifications under less dire conditions may indicate that the
servicer is actively managing the SPV’s assets. Active management would
in turn trigger a new layer of taxation on the SPV’s income, in addition to
the tax investors pay on their income from the RMBS.85 Moreover, because
81. See supra Part III.B.

82. See infra Part III.D.
83. If the SPV could file or be placed in bankruptcy, it could reject the PSA with the servicer as
an executory contract. See 11 U.S.C. § 365(a), (h). The servicer would receive only an unsecured
prepetition claim for damages, and the SPV would then be free to recontract for servicing under
nonrestrictive terms. All contracts may be breached of course, regardless of bankruptcy, but rejection of
a contract under § 365 of the Bankruptcy Code is treated as a prebankruptcy breach, rather than a
postbankruptcy breach, which changes the treatment of the counterparty’s claim. See id. § 365(g).
To reject a contract under § 365, the contract must be executory, a term not defined in the
Bankruptcy Code but commonly held to mean that material performance is due from both parties to the
contract such that “the failure of either to complete performance would constitute a material breach
excusing the performance of the other.” See Vern Countryman, Executory Contracts in Bankruptcy:
Part I, 57 MINN. L. REV. 439, 460 (1973). Because most PSAs are also part of the MBS indenture, there
is the question of whether the PSA is executory. While the servicer and the SPV have ongoing mutual
obligations, the MBS holders do not, and unless the PSA is severable, the nonexecutory nature of the
MBS holders’ relationship with the SPV might render the PSA nonexecutory and therefore not
rejectable.
84. Hunt, supra note 59, at 7.
85. Most RMBS are structured to qualify as real estate mortgage investment conduits
(“REMICs”) under the Internal Revenue Code. See 26 U.S.C. § 860A (2006) (REMIC tax treatment).
Generally, investors in a Subchapter C corporation are subject to double taxation—the corporation is
the originator is often the servicer, overly active management of the

securitized loans could suggest that it did not truly sell the risk. In response,
regulators could require the servicer/originator to bring the loans back onto
its balance sheet, defeating the point of securitization.86
taxed directly on its earnings, and then the investors are taxed on any distributions from the corporation.
REMICs, however, receive pass-through tax status, meaning that the investors, rather than the entity,
are taxed on the REMIC’s earnings. Id. § 860A(b). See also ACCOUNTING FOR TRANSFERS AND
SERVICING OF FIN. ASSETS AND EXTINGUISHMENTS OF LIABILITIES, Statement of Fin. Accounting
Standards No. 140 (Fin. Accounting Standards Bd. 2000) (off–balance sheet accounting treatment).
To qualify as a REMIC, the entity must be essentially passive in its management of mortgages.
SCHWARCZ ET AL., supra note 22, at 114. It is limited in its ability to change what mortgage loans it
owns, see 26 U.S.C. § 860D(a)(4) (limiting REMICs to holding “qualified mortgages” and “permitted
investments”); id. § 860G(a)(3) (defining “qualified mortgage”); id. § 860G(a)(5) (defining “permitted
investments”), or to alter the terms of the mortgage loans, see 26 C.F.R. § 1.860G-2(b) (2008). To
qualify as a REMIC under the Internal Revenue Code, substantially all of the REMIC’s assets must be
qualified mortgages and permitted investments. 26 U.S.C. § 860D(a)(4). “Qualified mortgage” is
defined as “any obligation (including any participation or certificate of beneficial ownership therein)
which is principally secured by an interest in real property.” Id. § 860G(a)(3). It includes “foreclosure
property,” see id., which is defined as property “which is acquired in connection with the default or
imminent default of a qualified mortgage held by the REMIC,” id. § 860G(a)(8)(B).
The Treasury regulations note, however, that if a mortgage is significantly modified, other than in
the event of a “default or reasonably foreseeable default,” then “the modified [mortgage] is treated as
one that was newly issued in exchange for the unmodified obligation that it replaced,” which means that
the modified mortgage will not be a qualified mortgage. 26 C.F.R. § 1.860G-2(b). The tax consequence
of this recharacterization is that the deemed disposition of the unmodified obligation will be a
prohibited transaction that is subject to a 100 percent tax. 26 U.S.C. § 860F(a).
The relevance of a mortgage modification depends on whether the modification is considered
“significant” under one of five specific categories. 26 C.F.R. § 1.1001-3(b), (e). Most relevant for
mortgage modification purposes is the specific category describing certain yield changes as significant
modifications. See id. § 1.1001-3(e)(2). A modification that results in a change in the yield on the
mortgage loan of greater of twenty-five basis points, or 5 percent of the annual unmodified yield, is
considered significant. Id. § 1.1001-3(e)(2)(ii).
If a modification does not fit into any of the specific significant modification categories, then
whether the modification alters legal rights or obligations in an economically significant manner must
be analyzed. Id. § 1.1001-3(e)(1). Moreover, all modifications are considered collectively. Id.
Two consequences flow from a modification being significant. First, the modified mortgage held
by the trust will not be a qualified mortgage. If a trust has too many nonqualified mortgages, the trust
loses its REMIC status. Second, the trust will incur 100 percent taxation on the net gain in the exchange
deemed to have occurred in the modification. Therefore, too many modifications without imminent
default will result in the loss of the SPV’s REMIC pass-through status, not just for specific transactions,
but overall. The Internal Revenue Service has relaxed application of REMIC rules to mortgage loan
modification programs where the mortgage is secured by owner-occupied property and there is a
significant risk of foreclosure. See Rev. Proc. 2008-28, 2008-23 I.R.B. 1054, 1054–55.
Because pass-through tax status is a crucial element of RMBS’ economic value, trusts are often
structured to ensure that REMIC status is protected, such as by specifically limiting servicers’ ability to
modify loans in ways that would endanger REMIC status.
86. Accounting concerns dovetail with tax exemption requirements: both require the trust to be
passively managed. The servicer is frequently the originator, and in order to ensure that the securitized
assets may be removed from the originator’s balance sheet and that the originator can recognize the
gain on the sale of the mortgage loans in its transaction with the SPV, the SPV must be “qualified,”
In sum, SPVs are structured as immortal automatons: they cannot go

bankrupt, and they are immune to management discretion. Once launched,
they are meant to operate more or less on autopilot until the securities they
issue are paid off. This effectively impedes renegotiation of the PSA and
the underlying assets.
D. FUNCTIONAL RIGIDITY: COORDINATION PROBLEMS FROM TRANCHING,

RESECURITIZATION, AND INSURANCE
Loan- and pool-level rigidities are reinforced through tranching,

resecuritization, and third-party insurance. Non-government-sponsored
RMBS issued by an SPV are typically tranched—divided into a stepped
senior/subordinated payment priority system where the subordinated
tranches are first in line to absorb losses from reduced mortgage
payments.87 Each tranche usually carries a different rate of return and a
different credit rating.88 Sometimes the tranching is done separately for
principal and interest payments due on the mortgages.89
Senior/subordinate tranching is a hallmark of private-label asset
securitization.90 This contrasts with corporate finance practice, where
which means inter alia that the originator must have no control over the assets. See ACCOUNTING FOR
TRANSFERS AND SERVICING OF FIN. ASSETS AND EXTINGUISHMENTS OF LIABILITIES, Statement of Fin.
Accounting Standards No. 140, ¶¶ 8–13 (Fin. Accounting Standards Bd. 2008), available at
http://www.fasb.org/pdf/aop_FAS140.pdf (amending the 2000 version). Financial accounting standards
do not specify what would constitute “control,” but the possibility of balance sheet consolidation makes
originator-servicers chary of active management of securitized mortgages, including modification of
their terms. Securities and Exchange Commission staff, however, have indicated that they do not
believe that modifications of imminently defaulting loans would require on–balance sheet accounting.
Letter from Christopher Cox, SEC Chairman, to Rep. Barney Frank, Chairman of the Comm. on Fin.
Servs., U.S. House of Representatives (July 24, 2007), available at
http://financialservices.house.gov/072407SEC.pdf; Letter from Conrad Hewitt, SEC Chief Accountant,
to Arnold Hanish, Chairman of the Comm. on Corporate Reporting, Fin. Executives Int’l, and Sam
Ranzilla, Chairman of the Prof’l Practice Executive Comm., Am. Inst. of Certified Pub. Accountants
(Jan. 8, 2008), available at http://www.sec.gov/info/accountants/staffletters/hanish010808.pdf.
87. See OCC HANDBOOK, supra note 24, at 19–20.
88. See id. See also FRANK J. FABOZZI, ANAND K. BHATTACHARYA & WILLIAM S. BERLINER,
MORTGAGE-BACKED SECURITIES: PRODUCTS, STRUCTURING, AND ANALYTICAL TECHNIQUES 23
(2007); Patricia A. McCoy, Andrey D. Pavlov & Susan M. Wachter, Systemic Risk Through
Securitization: The Result of Deregulation and Regulatory Failure, 41 CONN. L. REV. 1327, 1331–32
(2009).
89. See OCC HANDBOOK, supra note 24, at 24.
90. FABOZZI ET AL., supra note 88, at 22–24. Private-label securitization covers all securitization
other than by Fannie Mae, Freddie Mac, and Ginnie Mae. Fannie Mae, Freddie Mac, and Ginnie Mae
RMBS are not tranched for credit risk, as all GSE RMBS have a GSE credit guaranty and all Ginnie
Mae RMBS bear a federal guaranty. Some Fannie Mae, Freddie Mac, and Ginnie Mae collateralized
mortgage obligations (“CMOs”) are tranched for prepayment priority, however, as a way of allocating
interest rate risk.
bonds issued under a single indenture are not tranched. Senior/subordinate

tranching under a single indenture creates incentives against modification:
junior tranches, which stand to lose the most from reducing flows into the
SPV, may block modification or hold it up for ransom. In corporate bond
practice, where securities of different ranking are issued under different
indentures, junior creditors are only in a free-rider position: senior creditors
may restructure if they wish, even if the juniors will benefit from their
concessions. In contrast, the holders of a subordinated tranche of RMBS
have a potential veto over renegotiation of the PSA.
Consider a situation in which a renegotiation of the underlying
mortgages would reduce the aggregate payment streams to RMBS
investors, but less so than if there were no renegotiation. The modified
mortgage loans would be more valuable than if they defaulted and went
into foreclosure, but less valuable than if they performed as originally
intended. The loss from renegotiation would be borne first by the
subordinated tranches. The subordinated tranches have little incentive to
allow renegotiation of the underlying mortgages unless they receive some
of the benefit from it.91 If the subordinated tranches are “out of the money,”
they will not consent to the modification of the PSA unless they get a side
payment.
Likewise, the senior-most tranches may have no incentive to
cooperate in a renegotiation, as they incur none of the benefit from
modifying the mortgages—they will get paid no matter what and might
even get paid faster with foreclosure. All the benefit of renegotiation
accrues to the “fulcrum” tranche that is in the money if there is a
modification and out of the money in a foreclosure.92
The SPV has neither the funds nor the authority to pay off dissenting
trancheholders. As a result, if either senior or subordinate tranches’ consent
is required to modify the PSA or the underlying mortgages, renegotiation
will likely stall absent side payments from outside the securitization
structure.
Tranching also means that even if an SPV were not bankruptcy
remote, it is much less likely to end up in bankruptcy. Bankruptcy enforces
91. See Standard & Poor’s RatingsDirect, U.S. Housing’s Long and Winding Road to Recovery
8 (Apr. 14, 2009), available at http://www2.standardandpoors.com/spf/pdf/fixedincome/
US_Housing_Long.pdf.
92. This is related to David Skeel’s idea of a “pivotal” class in corporate bankruptcy. See David
Arthur Skeel, Jr., The Nature and Effect of Corporate Voting in Chapter 11 Reorganization Cases, 78
VA. L. REV. 461, 480 (1992).
subordination agreements to the extent they are enforceable outside of

bankruptcy.93 This reduces the value of bankruptcy for RMBS holders,
since the incentives described above continue unchanged in bankruptcy.
Senior/subordinate tranching substitutes contractual ordering of distress
(absolute priority) for the statutory ordering and judicial process of
bankruptcy. Unlike bankruptcy, which is heavily negotiated and can give
third parties (“parties in interest”) a voice in reorganization, contractual
loss allocation is locked in, along with the attendant incentive structure and
externalities.
More collective action problems arise through resecuritization.
Because the riskier tranches of an RMBS issuance are not investment
grade, they cannot be sold to entities like pension plans and mutual funds.
These tranches are often resecuritized, either into collateralized mortgage
obligations (“CMOs”) or collateralized debt obligations (“CDOs”).94
A CMO is a securitization backed by mortgage-backed securities
rather than by mortgages.95 A CDO is a securitization backed by a variable
pool of assets, potentially including mortgage-backed securities, as well as
other types of securitizations, bonds, and loan interests.96 CMOs and CDOs
are themselves then tranched.97 Senior CMO and CDO tranches can receive
investment-grade ratings because of the credit enhancement provided by
the subordinated tranches.98 Thus, it is possible to transform non-
investment-grade tranches of RMBS into investment-grade tranches of
CMOs and CDOs that can be sold to conservative institutional investors
who are restricted to purchasing investment-grade securities. The non-
investment-grade components of CMOs and CDOs can themselves be
resecuritized once again into what are known as CMO2s and CDO2s, with
the senior tranches of the CMO2s and CDO2s receiving investment-grade
ratings.99 This process can be repeated an endless number of times. The
result is impressive. The CEO of Goldman Sachs observed that “[i]n
January 2008, there were 12 triple A–rated companies in the world. At the
same time, there were 64,000 structured finance instruments . . . rated triple
93. 11 U.S.C. § 510(a) (2006).

94. FABOZZI ET AL., supra note 88, at 23–24.
95. Id. at 23.
96. See McCoy et al., supra note 88, at 1331. CDOs are often actively managed. Jennifer E.
Bethel, Allen Ferrell & Gang Hu, Legal and Economic Issues in Litigation Arising from the 2007–2008
Credit Crisis 10–11 (John M. Olin Ctr. for Law, Econ. & Bus., Harvard Law Sch., Faculty Discussion
Paper No. 612, 2008), available at http://ssrn.com/abstract=1096582.
97. See McCoy et al., supra note 88, at 1331.
98. See id.
99. FABOZZI ET AL., supra note 88, at 23; McCoy et al., supra note 88, at 1331.
A.”100 Securitization has turned dross into gold, or at least non-investment-

grade dross into AAA-rated dross. See Figure 4 for an illustration of how
mezzanine tranches of RMBS deals are resecuritized into a CDO.
FIGURE 4. Resecuritization of RMBS into a CDO
The upshot of this financial alchemy is that to control 100 percent of

an RMBS issuance in order to alter a PSA, one would also have to own 100
percent of multiple CMOs and CDOs to alter the CMOs’ and CDOs’ PSAs,
and of multiple CMO2s and CDO2s to alter the CMO2s’ and CDO2s’ PSAs.
This process can occur separately for multiple mortgages on the same
property. When amending the underlying mortgages requires a vote,
resecuritized interests similarly must be taken into account. Securing a
supermajority vote is only slightly less complicated.
The final obstacle to obtaining consent for PSA and mortgage
modification combines elements of resecuritization and insurance. The
100. Lloyd Blankfein, Do Not Destroy the Essential Catalyst of Risk, FIN. TIMES (London), Feb.
8, 2009, at 7.
following example illustrates. An SPV’s income can exceed the coupons it

must pay certificate holders. For example, the SPV may be
overcollateralized, so that it holds 10,000 mortgages but only needs to hold
9500 performing mortgages to meet its bond coupon obligations.
Alternatively, the coupons might be fixed rate, but the mortgages might be
adjustable rate, so if the adjustable rate exceeds the coupon rate, the SPV
will retain the difference. In such cases, the SPV can have substantial
residual value after all investors are paid.
The residual value of the SPV after the certificate holders are paid is
called the net interest margin (“NIM”).101 The NIM is typically
resecuritized separately into a NIM security (“NIMS”),102 and the NIMS is
often insured by a financial institution. NIMS insurers’ consent is typically
required both to modify PSAs and to modify the underlying mortgages
beyond limited thresholds.103 NIMS insurers hold positions similar to
equity holders for SPVs. Even more so than junior trancheholders in the
very first securitization, NIMS insurers have little incentive to cooperate.
They have nothing to lose, but there is no money or authority for the SPV
to pay them off.
E. SUMMARY: RIGID BY DESIGN
In sum, securitization creates multiple kinds and layers of contractual

rigidity that prevent the renegotiation of RMBS and the underlying assets,
even when renegotiation would be the welfare-maximizing outcome.
Securitization design farms out and wills away credit risk through a
combination of explicit contractual commitment, legal structure or
organizational form, and a financial structure prone to coordination
problems. As part of this design, tax treatment, accounting, tranching, and
insurance have the powerful effect of reinforcing commitment to pay, even
though that is not their primary purpose.
It bears emphasis that layered rigidity is not an accidental byproduct
of securitization design; for private-label residential mortgage
securitization, at least, layered rigidity is among its central premises. This
is so even where rigidity is not the primary goal of a particular design
feature (such as tax treatment), and where real-life securitization contracts
may fail to achieve perfect immutability. This can be seen from a
comparison of private-label RMBS with commercial mortgage-backed
101. FABOZZI ET AL., supra note 88, at 199.

102. Id.
103. Id.
securities (“CMBS”).
CMBS are structured very differently from RMBS. In particular, they
are designed with the need for workouts in mind. There is greater
heterogeneity among CMBS PSAs than RMBS PSAs, but CMBS PSAs
commonly contain several flexibility-enhancing features that are rare
among, or entirely absent from, RMBS. First, CMBS usually feature a
special default servicer that assumes management of any loans that are
sixty or more days delinquent.104 CMBS special servicers are subject to
many fewer restrictions than private-label RMBS servicers. They are often
explicitly authorized to do principal reductions, interest rate reductions,
reamortizations, term stretchouts, and temporary forbearance.105
104. CMBS servicers cannot generally modify performing loans without risking adverse tax
consequences for the trust. See Lingling Wei, New Rules Ease the Restructuring of CMBS Loans, WALL
ST. J., Sept. 16, 2009, at C6 (noting that the U.S. Department of Treasury had relaxed some tax
provisions in light of such adverse consequences).
105. See, e.g., Bear Stearns Commercial Mortgage Securities Trust 2002-TOP6, Pooling and
Servicing Agreement (Form 8-K), at EX-4.1 § 9.5(c) (Apr. 3, 2002), available at
http://www.secinfo.com/$/SEC/Registrant.asp?CIK=1168574 (“Subject to the Servicing Standard and
Sections 9.39 and 9.40, and the rights and duties of the Master Servicer under Section 8.18, the Special
Servicer may enter into any modification, waiver or amendment (including, without limitation, the
substitution or release of collateral or the pledge of additional collateral) of the terms of any Specially
Serviced Mortgage Loan, including any modification, waiver or amendment to (i) reduce the amounts
owing under any Specially Serviced Mortgage Loan by forgiving principal, accrued interest and/or any
Prepayment Premium, (ii) reduce the amount of the Scheduled Payment on any Specially Serviced
Mortgage Loan, including by way of a reduction in the related Mortgage Rate, (iii) forbear in the
enforcement of any right granted under any Mortgage Note or Mortgage relating to a Specially Serviced
Mortgage Loan, (iv) extend the Maturity Date of any Specially Serviced Mortgage Loan and/or
(v) accept a principal prepayment on any Specially Serviced Mortgage Loan during any period during
which voluntary Principal Prepayments are prohibited, provided, in the case of any such modification,
waiver or amendment, that (A) the related Mortgagor is in default with respect to the Specially Serviced
Mortgage Loan or, in the reasonable judgment of the Special Servicer, such default is reasonably
foreseeable, (B) in the reasonable judgment of the Special Servicer, such modification, waiver or
amendment would increase the recovery on the Specially Serviced Mortgage Loan to Certificateholders
on a net present value basis (the relevant discounting of amounts that will be distributable to
Certificateholders to be performed at related Mortgage Rate) . . . . In no event, however, shall the
Special Servicer (i) extend the Maturity Date of a Specially Serviced Mortgage Loan beyond a date that
is two years prior to the Rated Final Distribution Date . . . .”); Credit Suisse Commercial Mortgage
Trust/Series 2007-C1, Pooling and Servicing Agreement (Form 8-K), at EX-4.1 § 3.20(a) (Apr. 2,
2007), available at http://www.secinfo.com/dRSm6.us2.c.htm#1stPage [hereinafter Credit Suisse 2007-
C1 PSA] (“The Special Servicer (solely as to the Specially Serviced Mortgage Loans) or the Master
Servicer (solely as to the Performing Mortgage Loans) may (consistent with the Servicing Standard)
agree to any modification, waiver or amendment of any term of, extend the maturity of, defer or forgive
interest (including Default Interest and Post-ARD Additional Interest) on and principal of, defer or
forgive late payment charges and Yield Maintenance Charges on, permit the release, addition or
substitution of collateral securing, and/or permit the release, addition or substitution of the Borrower on
or any guarantor of, any Mortgage Loan . . . .”); Credit Suisse Commercial Mortgage Trust Series 2007-
C5, Pooling and Servicing Agreement (Form 8-K), at EX-4.1 § 3.20(d) (Nov. 27, 2007), available at
http://www.secinfo.com/dRSm6.u2fr.c.htm#ggyk [hereinafter Credit Suisse 2007-C5 PSA]
Second, compensation for CMBS special default servicers incentivizes

them to perform workouts that are economically beneficial for the CMBS
investors as a whole. CMBS special default servicers are paid an annual
percentage of the principal for all loans being serviced (that is, loans in
default).106 If the loan is worked out and begins to perform, however, the
CMBS special servicer then receives a higher percentage fee of every
payment made of principal and interest.107 This gives the CMBS special
servicer an incentive to get the loan performing again, but also not to
modify it more than necessary, as its compensation is based on the
payments on the modified loan.
CMBS PSAs also provide that special servicers are reimbursed with
interest (at prime) for servicing advances.108 This contrasts with RMBS
(“Notwithstanding Sections 3.20(a)(ii) and 3.20(c), but subject to Sections 3.20(e) and 3.20(f), the
Special Servicer may (i) reduce the amounts owing under any Specially Serviced Mortgage Loan by
forgiving principal, accrued interest and/or any Yield Maintenance Charge or Static Prepayment
Premium, (ii) reduce the amount of the Monthly Payment on any Specially Serviced Mortgage Loan,
including by way of a reduction in the related Mortgage Rate, (iii) forbear in the enforcement of any
right granted under any Note or Mortgage relating to a Specially Serviced Mortgage Loan, (iv) extend
the maturity of any Specially Serviced Mortgage Loan, (v) waive Excess Interest if such waiver
conforms to the Servicing Standard, (vi) permit the release or substitution of collateral for a Specially
Serviced Mortgage Loan and/or (vii) waive a Yield Maintenance Charge or Static Prepayment Premium
or accept a Principal Prepayment during any lockout period; provided that (A) the related Borrower is
in default with respect to the Specially Serviced Mortgage Loan or, in the judgment of the Special
Servicer, such default is reasonably foreseeable and (B) in the sole good faith judgment of the Special
Servicer and in accordance with the Servicing Standard, such modification would result in recovery that
equals or exceeds recovery for liquidation on the subject Mortgage Loan to Certificateholders . . . as a
collective whole, on a present value basis (the relevant discounting of amounts that will be distributable
to Certificateholders or a B Loan Holder or Companion Loan Holder to be performed at a rate not less
than the related Mortgage Rate).”).
CMBS PSAs will typically place limits on term stretchouts. Like RMBS PSAs, they limit term
stretchouts to ensure that the trusts’ revenues will have durations that match the trusts’ obligations.
106. See, e.g., Credit Suisse 2007-C5 PSA, supra note 105, § 3.11(b); id. § 1.01 (defining “Special
Servicing Fee” and “Special Servicing Fee Rate” as thirty-five basis points per annum per loan, with a
minimum $4000 per month, reducible by the Directing Certificateholder.).
107. See, e.g., id. § 3.11(b) (“The Special Servicer shall also be entitled to additional servicing
compensation in the form of a Workout Fee with respect to each Corrected Mortgage Loan at the
Workout Fee Rate. The Workout Fee shall be payable out of, and shall be calculated by application of
the Workout Fee Rate to, each collection of interest, other than Default Interest and Excess Interest, and
principal . . . received on such Mortgage Loan for so long as it remains a Corrected Mortgage Loan. The
Workout Fee with respect to any Corrected Mortgage Loan will cease to be payable if such Mortgage
Loan again becomes a Specially Serviced Mortgage Loan; provided that a new Workout Fee will
become payable if and when such Mortgage Loan again becomes a Corrected Mortgage Loan.”); id.
§ 1.01 (defining “Workout Fee” and “Workout Fee Rate” as 1 percent).
108. See, e.g., id. § 3.03(d) (“In connection with its recovery of any Servicing Advance . . . [the]
Servicer, the Special Servicer and the Trustee shall each be entitled to receive . . . interest at the
Reimbursement Rate in effect from time to time, accrued on the amount of such Servicing Advance
from and including the date made to, but not including, the date of reimbursement . . . .”); id. § 1.01
PSAs, which do not provide for servicers to be reimbursed for the interest
expense of servicing advances. Compensation for interest expense means
that special servicers will not rush either to overmodify a loan so it will
perform, which would terminate the obligation to make servicing advances,
or to foreclose and liquidate the property, which would also end the duty to
make advances.
A third flexibility-enhancing feature of CMBS is a mechanism that
vests a single investor with decisionmaking and monitoring capability. In
CMBS deals, there is either a controlling party or controlling class
representative entitled to disclosures from the servicer and trustee and
empowered to direct or veto a variety of trust management decisions. Some
CMBS PSAs feature a controlling party, referred to as the “Directing
Certificateholder” or “Controlling Certificateholder,” who is the holder of
the majority of the residual class of claims.109 Other CMBS PSAs permit
the residual class of claims to elect a “Controlling Class Representative.”110
Critically, the identity of the controlling party shifts with the
performance of the trust’s assets because the residual class is not a fixed
class. Instead, it adjusts according to where the cashflow waterfall stops at
(defining “Reimbursement Rate” as “the Prime Rate”); Morgan Stanley Capital I Trust Commercial
Mortgage 2007-Top25, Pooling and Servicing Agreement (Form 8-K), at EX-4.1 § 4.5 (Feb. 14, 2007)
(“Any unreimbursed Advance funded from the Master Servicer's, the Special Servicer’s or the Trustee’s
own funds shall accrue interest on a daily basis, at a per annum rate equal to the Advance Rate, from
and including the date such Advance was made to but not including the date on which such Advance
has been reimbursed . . . .”); id. § 1.1 (defining “Advance Rate” as “a per annum rate equal to the Prime
Rate as published in the ‘Money Rates’ section of The Wall Street Journal from time to time or such
other publication as determined by the Trustee in its reasonable discretion”).
109. There is variation in PSA terminology for this controlling shareholder.
110. See, e.g., Credit Suisse 2007-C5 PSA, supra note 105, § 1.01 (defining “Controlling Class”
to be, “[a]s of any date of determination, the Class of Principal Balance Certificates with the lowest
payment priority pursuant to Sections 4.01(a) and 4.01(b), that has a then outstanding Class Principal
Balance that is not less than 25% of its initial Class Principal Balance; provided that, if no Class of
Principal Balance Certificates has a Class Principal Balance that satisfies the foregoing requirement,
then the Controlling Class shall be the Class of Principal Balance Certificates with the lowest payment
priority pursuant to Sections 4.01(a) and 4.01(b), that has a then outstanding Class Principal Balance
greater than zero . . . [and] [a]s of the Closing Date, the Controlling Class shall be the Class S
Certificates”); id. (defining the “Series 2007-C5 Directing Certificateholder” as “[t]he particular
Holder . . . of Certificates of the Controlling Class selected by the Holders . . . of Certificates
representing more than 50% of the Percentage Interests in the Controlling Class . . . ; provided,
however, that until a Series 2007-C5 Directing Certificateholder is so selected or after receipt of a
notice from the Holders . . . of Certificates representing more than 50% of the Percentage Interests in
the Controlling Class that a Series 2007-C5 Directing Certificateholder is no longer designated, the
particular Certificateholder . . . that beneficially owns Certificates of the Controlling Class that
represents the largest aggregate Percentage Interest in the Controlling Class shall be the Series 2007-C5
Directing Certificateholder”).
any point in time.111 Thus, as the junior-most tranches find themselves out
of the money, control shifts upward in the capital structure. This means that
out-of-the-money junior tranches therefore have no say over decisions that
will no longer impact them. Likewise senior, well-in-the-money tranches
also have no say over decisions from which they are insulated by virtue of
still-in-the-money subordinated tranches. Instead, the CMBS controlling
party system means that an investor with money immediately on the line is
involved with management of the trust’s assets. CMBS PSAs recognize
that the interests of the controlling party might conflict with those of other
investors, and explicitly waive claims against the controlling party except
for actions taken negligently or in bath faith or for willful misfeasance.112
The CMBS controlling party is entitled to special information
disclosures from the servicer and trustee.113 The CMBS controlling party is
111. See, e.g., Credit Suisse 2007-C1 PSA, supra note 105, § 1.01 (“‘Controlling Class’ shall
mean, as of any date of determination, the eligible Class of Principal Balance Certificates with the
lowest payment priority pursuant to Sections 4.01(a) and 4.01(b), that has a then outstanding Class
Principal Balance that is not less than 25% of its initial Class Principal Balance; provided that, if no
eligible Class of Principal Balance Certificates has a Class Principal Balance that satisfies the foregoing
requirement, then the Controlling Class shall be the eligible Class of Principal Balance Certificates
with the lowest payment priority pursuant to Sections 4.01(a) and 4.01(b), that has a then outstanding
Class Principal Balance greater than zero. . . . As of the Closing Date, the Controlling Class shall be the
Class T Certificates.”); id. § 3.23(a) (“The Holders . . . of Certificates representing more than 50% of
the Class Principal Balance of the Controlling Class shall be entitled in accordance with this Section
3.23 to select a representative (each, a ‘Controlling Class Representative’) having the rights and powers
specified in this Agreement (including those specified in Section 3.24) or to replace an existing
Controlling Class Representative.”).
112. See, e.g., id. § 3.24(c) (“Each Certificateholder acknowledges and agrees, by its acceptance
of its Certificates, that: (i) the Controlling Class Representative may have special relationships and
interests that conflict with those of Holders of one or more Classes of Certificates; (ii) the Controlling
Class Representative may act solely in the interests of the Holders of the Controlling Class; (iii) the
Controlling Class Representative does not have any duties to the Holders of any Class of Certificates
other than the Controlling Class (and with respect to such Controlling Class Holders shall have no
liability for any action taken or omitted which does not constitute negligence, bad faith or willful
misfeasance); (iv) the Controlling Class Representative may take actions that favor interests of the
Holders of the Controlling Class over the interests of the Holders of one or more other Classes of
Certificates; and (v) the Controlling Class Representative shall have no liability whatsoever for having
so acted or for any action taken or omitted, and no Certificateholder may take any action whatsoever
against the Controlling Class Representative or any director, officer, employee, agent or principal
thereof for having so acted.”).
113. See, e.g., id. § 3.24(a) (“Upon reasonable request, the Special Servicer shall provide the
Controlling Class Representative with any information in such Special Servicer's possession with
respect to such matters [as foreclosure and modification and real estate–owned sales], including,
without limitation, its reasons for determining to take a proposed action. The Master Servicer or the
Special Servicer, as applicable, shall notify the Controlling Class Representative of any release or
substitution of collateral for a Mortgage Loan even if such release or substitution is in accordance with
the related Mortgage Loan Documents.”); id. § 11.10 (“The Trustee, the Master Servicer or the Special
Servicer, as the case may be, shall deliver to the Controlling Class Representative a copy of each notice
also empowered to direct or veto a variety of trust management decisions,

including foreclosures, modifications, amendment or waivers of monetary
terms and material nonmonetary terms, settlements with mortgagors, and
real estate–owned (“REO”) sales.114 While CMBS servicers have authority
or other item of information such Person is required to deliver to the Rating Agencies pursuant to
Section 11.09, in each case simultaneously with the delivery thereof to the Rating Agencies, to the
extent not already delivered pursuant to this Agreement.”).
114. See, e.g., id. § 3.24(a) (“The Controlling Class Representative will be entitled to advise the
Special Servicer with respect to the Special Servicer’s taking, or consenting to the Master Servicer’s
taking, any of the actions identified in clauses (i) through (x) of the following sentence. In addition,
notwithstanding anything in any other Section of this Agreement to the contrary, but in all cases subject
to Section 3.20(g) and Section 3.24(b), the Special Servicer will not be permitted to take, or consent to
the Master Servicer’s taking, any of the actions identified in clauses (i) through (x) of this sentence,
unless and until such Special Servicer has notified the Controlling Class Representative in writing of
such Special Servicer’s intent to take or permit the particular action and the Controlling Class
Representative has consented (or has failed to object) thereto in writing within five Business Days of
having been notified thereof in writing and having been provided with all reasonably requested
information with respect thereto: (i) any proposed foreclosure upon or comparable conversion (which
may include acquisitions of an REO Property) of the ownership of the property or properties securing
any Specially Serviced Mortgage Loans as come into and continue in default; (ii) any modification,
amendment or waiver of a monetary term (including any change in the timing of payments but
excluding the waiver of Default Charges) or any material non-monetary term (excluding any waiver of
a ‘due-on-sale’ or ‘due-on-encumbrance’ clause, which clauses are addressed in clause (ix) below) of a
Mortgage Loan; (iii) any acceptance of a discounted payoff with respect to any Specially Serviced
Mortgage Loan; (iv) any proposed sale of an REO Property for less than the Stated Principal Balance
of, and accrued interest (other than Default Interest and Post-ARD Additional Interest) on, the related
Mortgage Loan, except in connection with a termination of the Trust Fund pursuant to Section 9.01;
(v) any determination to bring an REO Property into compliance with applicable environmental laws or
to otherwise address Hazardous Materials located at an REO Property; (vi) any release of collateral for
any Mortgage Loan (other than in accordance with the specific terms which do not provide for lender
discretion of, or upon satisfaction of, such Mortgage Loan); (vii) any acceptance of substitute or
additional collateral for any Specially Serviced Mortgage Loan (other than in accordance with the
specific terms of such Mortgage Loan); (viii) any release (other than in accordance with the related
Mortgage Loan Documents and in an amount less than $50,000) of Earn-Out Reserve Funds or related
Letter of Credit with respect to a Mortgaged Property securing a Mortgage Loan; (ix) any waiver of a
due-on-sale or due-on-encumbrance clause in any Mortgage Loan; and (x) any consent to a change in
franchise with respect to a hospitality loan or a change in the property manager of a Mortgage Loan
with a principal balance greater than $5,000,000; provided that, if the Special Servicer or the Master
Servicer, as applicable, determines that immediate action is necessary to protect the interests of the
Certificateholders (as a whole), the Special Servicer or the Master Servicer, as the case may be, may
take any such action without waiting for the response of the Controlling Class Representative to the
Special Servicer. In addition, subject to Section 3.24(b), the Controlling Class Representative may
direct the Special Servicer to take, or to refrain from taking, such actions as such Controlling Class
Representative may deem advisable or as to which provision is otherwise made herein. Upon reasonable
request, the Special Servicer shall provide the Controlling Class Representative with any information in
such Special Servicer's possession with respect to such matters, including, without limitation, its
reasons for determining to take a proposed action. The Master Servicer or the Special Servicer, as
applicable, shall notify the Controlling Class Representative of any release or substitution of collateral
for a Mortgage Loan even if such release or substitution is in accordance with the related Mortgage
Loan Documents.”).
to act on their own, they are subject to direction of the controlling party.115
CMBS controlling parties also have a powerful disciplinary tool at their
disposal to ensure special servicer cooperation: CMBS controlling parties
may fire the special servicer without cause.116 In contrast, RMBS servicers
may be dismissed only for very limited, specified causes. In RMBS, the
effective principals (the RMBS investors) have little ability to discipline
servicers; at best they are disciplined reputationally, but in an industry as
opaque as servicing, reputation is at most weak market discipline.117 The
dismissal power means that CMBS special servicers are likely to be attuned
to the interests of the residual tranche. CMBS PSAs thus create a structure
that ensures better principal-agent interest alignment and control.
CMBS PSAs show that it is possible to have a much more flexible
securitization structure than what prevails for RMBS. Because
securitization need not be inherently inflexible, we must therefore ask why
RMBS PSAs are more rigid than CMBS PSAs. Part of the answer may lie
in the nature of the securitized assets. CMBS typically have many fewer
underlying properties. Instead of thousands or tens of thousands of
mortgage loans in RMBS, there are typically dozens or hundreds in CMBS.
115. See, e.g., id. § 3.24(b) (“Notwithstanding anything herein to the contrary, (i) the Special
Servicer shall not have any right or obligation to consult with or to seek and/or obtain consent or
approval from the Controlling Class Representative prior to acting . . . and (ii) no advice, direction or
objection from or by the Controlling Class Representative, as contemplated by Section 3.24(a) or any
other provision of this Agreement, may (and the Master Servicer and Special Servicer shall ignore and
act without regard to any such advice, direction or objection that such Master Servicer or Special
Servicer, as the case may be, has determined, in its reasonable judgment, would) (A) require or cause
the Master Servicer, such Special Servicer or the Trustee to violate applicable law, the terms of any
Mortgage Loan or any other Section of this Agreement, including such Master Servicer’s or Special
Servicer’s obligation to act in accordance with the Servicing Standard, (B) result in an Adverse REMIC
Event with respect to either Trust REMIC or an Adverse Grantor Trust Event with respect to the
Grantor Trust, (C) expose the Trust, the Depositor, the Master Servicer, the Special Servicer, the
Trustee, or any of their respective Affiliates, members, managers, officers, directors, employees or
agents, to any material claim, suit or liability, or (D) expand the scope of the Master Servicer’s or
Special Servicer’s responsibilities under this Agreement.”).
116. See, e.g., id. § 3.25(a) (“Subject to Section 3.25(b), the Controlling Class Representative
may, upon not less than ten days’ prior written notice to the respective parties hereto, remove any
existing Special Servicer hereunder (with or without cause) and appoint a successor Special Servicer;
provided that, if any such removal is made without cause, then the costs of transferring the special
servicing responsibilities to a successor Special Servicer will, upon such removal or other termination,
be paid by the Certificateholders of the Controlling Class.”); Credit Suisse 2007-C5 PSA, supra note
105, § 7.01(d) (“The Series 2007-C5 Directing Certificateholder shall be entitled to terminate the rights
and obligations of the Special Servicer under this Agreement, with or without cause, upon 10 Business
Days prior written notice to the Master Servicers, the Special Servicer and the Trustee, and to appoint a
successor Special Servicer . . . .”).
117. See Levitin & Twomey, supra note 20, at 68–69 (discussing the limitations of reputational
discipline for RMBS servicers).
The loans in a CMBS pool are much larger, though, than the loans in an
RMBS pool. Loans in CMBS pools are also much more heterogeneous,
featuring individually negotiated loans with unique covenants and terms.
While private-label RMBS feature many so-called exotic home mortgage
products, there are a limited number of flavors of these exotic products;
they have nowhere near the diversity of commercial loan structures.
Because CMBS have many fewer loans, but for much larger amounts,
a single default is much more costly to CMBS holders than it is for RMBS
holders. Therefore, if x percent of the loans in a CMBS pool default, it will
have a much greater effect than if x percent of the loans in an RMBS pool
default. Lesser collateral diversification means that, all else being equal,
CMBS holders are more sensitive to default risk and therefore will value
loss mitigation more than RMBS holders. This factor weighs in favor of
having more flexible workout possibilities in CMBS PSAs.
Another factor potentially explaining the greater flexibility of CMBS
structures is the fact that commercial real estate market trends have been
more cyclical than residential real estate trends. There have been regional
residential market downturns in recent memory, but never a national
housing depression since the Great Depression. Accordingly, RMBS
investors were less likely to anticipate the need for loss mitigation
flexibility than CMBS investors, and did not demand it. Indeed, if RMBS
investors had anticipated the default rates between 2007 and the present,
they might not have purchased RMBS at all.
Ultimately, the reason that CMBS have more flexible structures than
RMBS appears to be because loss mitigation is expensive. The CMBS
special servicing structure adds another party and thus cost. The success fee
paid for reperforming loans, and the interest paid on servicing advances in
CMBS, reduces the yield available for CMBS investors. Moreover, the
special servicer discretion and the control party feature in CMBS adds
uncertainty to the investment. It means that CMBS are less likely to be
automatons than RMBS. While there are benefits to discretion, it can also
impose costs. CMBS thus feature a level of agency risk that RMBS do not,
but this also facilitates workouts. The nature of CMBS collateral means
that CMBS investors are more likely to demand flexible structures than
RMBS investors; ex ante, RMBS underwriters and investors do not tend to
think that flexibility is a good value proposition.
The comparison with CMBS shows that the hyperrigidity of RMBS is
not the inevitable byproduct of securitization. Instead, it is the product of
conscious design tradeoffs driven by cost considerations. RMBS investors
want greater yield, and RMBS sellers want higher prices. Loss mitigation is
expensive, and neither RMBS investors nor sellers want to pay for it. The
result is externalization of some of the costs of rigidity on homeowners,
communities, and the market, with the others borne by RMBS investors
themselves.
It is difficult to predict how these rigidities will work in practice to
constrain mortgage and PSA modification absent policy intervention.
Courts might interpret formal constraints liberally,118 pervasive distress
may create new financial and reputational pressures,119 and informal
coordination mechanisms may spring up to overcome the obstacles built
into the formal framework. Recent litigation120 and crisis experience so far
suggest that design rigidities in securitization do in fact work to impede
modification to some extent.121
The core argument of this part has been that it is impossible to
118. Cf. Hunt, supra note 59, at 10–12 (recommending the implementation of certain measures
that would facilitate interpreting PSAs as allowing mortgage loan modification).
119. This appears to have been the case with Ocwen Financial, a major subprime servicer that
began doing principal modifications of defaulted mortgages it serviced without regard to PSAs. See
Kate Berry, Debt Forgiveness: Ocwen Enters Uncharted Waters, AM. BANKER, June 24, 2008, at 1.
Ocwen lacked a captive funding source, so foreclosures were very expensive for it, as it was obligated
to make servicing advances of principal and interest on foreclosed mortgages until it realized funds on
the property. Servicing advances are recoverable, but without interest, so the time value can place a
heavy strain on a servicer’s liquidity.
120. See, e.g., Greenwich Complaint, supra note 56, ¶ 3 (“The . . . object of this action is a
declaration that, under the substantially identical agreements that govern the trust that sold the securities
owned by plaintiffs and the 373 other trusts in the CWL and CWALT securitizations that sold the
securities owned or held by other members of the plaintiff class, Countrywide is required to purchase
any loan on which it agrees to reduce the payments [as required by the terms of the PSA].”).
121. See Posting of Joe Nocera to N.Y. Times Executive Suite, http://
executivesuite.blogs.nytimes.com/2008/11/18/what-securitization-problem-the-fdic-weighs-in/ (Nov.
18, 2008, 18:45 EST) (citing both industry arguments that securitization contracts preclude
modification and the FDIC’s contrary interpretation). At a minimum, this suggests that the contractual
framework was designed to be rigid and is perceived as such by the relevant market actors, and that the
FDIC feels less legally and financially constrained than private actors to test the limits of the
contractual framework. See COMPTROLLER OF CURRENCY, U.S. DEP’T OF TREASURY, OCC AND OTS
MORTGAGE METRICS REPORT: DISCLOSURE OF NATIONAL BANK AND FEDERAL THRIFT MORTGAGE
LOAN DATA, FOURTH QUARTER 2008, at 23 (2009), available at http://www.occ.treas.gov/ftp/
release/2009-37a.pdf (finding significantly higher redefault rates for securitized loans than for other
loans and attributing the difference to the decreased flexibility of securitized loans); Alan M. White,
Rewriting Contracts, Wholesale: Data on Voluntary Mortgage Modifications from 2007 and 2008
Remittance Reports, 36 FORDHAM URB. L.J. 509, 525–29 (2009) (interpreting modification data to
suggest virtually no reduction in payment flows to securitization vehicles); Tomasz Piskorski, Amit
Seru & Vikrant Vig, Securitization and Distressed Loan Renegotiation: Evidence from the Subprime
Mortgage Crisis 1–2, 21–22 (Univ. of Chi. Booth Sch. of Bus., Working Paper No. 09-02, 2008),
available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1321646 (finding less renegotiation
and more foreclosures among securitized loans than among those held directly by the lender).
determine the actual rigidity of a contract by looking only, or even

principally, at “What . . . Subprime Securitization Contracts Actually Say
About Loan Modification.”122 To be sure, contract language is important.
Yet it is not determinative and can be misleading. Rigidity is a product not
just of the formal contractual limits on modification but also of the identity
and structure of transaction participants (including insurers and guarantors),
market structures (including investor dispersion), and contracting practices
(such as resecuritization). A covenant prohibiting amendment may be
ineffective where alternative debt restructuring techniques are available,123
while a mere supermajority approval requirement may be insurmountable
where investors are numerous, far flung, and lack incentives to modify.
The role of securitization and PSAs as barriers to mortgage
modification will no doubt evolve with experience, including litigation.124
Establishing the relationship between the incidence of modification and the
factors we have described, along with many other factors that might turn
out to be relevant, will take empirical work of the sort that is only
beginning to appear at this writing.125 For our purposes, it is enough that
the contractual arrangements we study are designed to be rigid and that
122. See Hunt, supra note 59. Hunt’s empirical study of PSA language is extremely valuable, but
the question it explores does not cover the full range of contract rigidity.
123. See, e.g., Lee C. Buchheit, How Ecuador Escaped the Brady Bond Trap, INT’L FIN. L. REV.,
Dec. 2000, at 17 (describing a debt restructuring in the face of an explicit prohibition and a unanimous
consent requirement for modification).
124. E.g., Greenwich Complaint, supra note 56. To date, litigation over improper servicing action
has not materialized, excluding the Greenwich Financial Services case, which was predicated on
Countrywide’s unique PSAs, as well as Countrywide’s settlement with the California and Illinois
attorneys general, in which Countrywide pledged to modify mortgages it had securitized and was
servicing. See, e.g., Stipulated Judgment and Injunction ¶ 6.3.1–3, People v. Countrywide Fin. Corp.,
No. LC083076 (Cal. Super. Ct. Oct. 20, 2008), available at http://ag.ca.gov/cms_attachments/press/
pdfs/n1618_cw_judgment.pdf. See also David Greising, Deal to Help 21,000 in State Keep Homes,
CHI. TRIB., Oct. 6, 2008, at C1.
125. See, e.g., Manuel Adelino, Kristopher Gerardi & Paul S. Willen, Why Don't Lenders
Renegotiate More Home Mortgages? Redefaults, Self-Cures, and Securitization (Fed. Reserve Bank of
Atlanta, Working Paper No. 2009-17, 2009), available at http://papers.ssrn.com/sol3/
papers.cfm?abstract_id=1433777 (comparing the incidence of modification between securitized and
portfolio loans and arguing that redefault risk, not securitization, is responsible for low rates of
modification). The authors also note the rarity of explicit modification bans in securitization contracts.
Id. at 24. On the other hand, recent government data suggest a dramatic difference in the type of
modification effected by portfolio and investor-held mortgages: the latter accounted for eight out of
17,574 principal reductions in the first two quarters of 2009, despite there being more than twice as
many investor-held mortgages modified (223,274 compared to 104,244). See Cong. Oversight Panel,
October Oversight Report: An Assessment of Foreclosure Mitigation Efforts After Six Months 56, 57
fig.21, 59 fig.23 (Oct. 9, 2009), available at http://cop.senate.gov/documents/cop-100909-report.pdf.
Similarly, term extensions are rare for private-label securitization mortgage modifications, but not for
loans in portfolio. Id. at 60 fig.24, 61 fig.25. Relating this research and data to the underlying law and
institutions is a rich subject for further empirical study.
they are perceived as such by key constituents, including investors,

servicers, and borrowers.126
Yet rigidity alone—or iron-clad commitment to pay—does not beget
capacity to pay. Contractual discipline of the sort described in the
preceding passages may fortify weak-willed debtors and servicers, but it
does not create money where there is none. There are limits to the alchemy.
The next two parts put PSAs in theoretical context and further consider the
implications of PSA design for managing widespread financial distress.
IV. THE GENUS: RMBS AS IMMUTABLE CONTRACTS
A. FORM, STRUCTURE, AND FUNCTION IN CONTRACT IMMUTABILITY
Contract doctrine disfavors modification bans. Justice Cardozo’s

pronouncement—“Those who make a contract, may unmake it. The clause
which forbids a change, may be changed like any other”127—remains for
the most part an accurate description of judicial attitudes. In contrast, some
modern contract theory has good things to say about immutable contracts.
Letting parties forswear even sensible future renegotiation is a common
theoretical response to information and incentive problems in contracting:
it can encourage disclosure and optimal investment up front and discourage
holdup behavior midperformance.128 It can also help minimize agency
costs.129
One way to achieve immutability is with express, formal amendment
restrictions, if they were enforced by the courts. The benefits of such
restrictions are most apparent in bilateral contracts. In multilateral
contracts—such as bonded debt—the benefits become attenuated because
these contracts are hard to modify even without express commitment to that
effect, owing to the extensive coordination efforts that are required to bring
126. See supra notes 120–21, 125.

127. Beatty v. Guggenheim Exploration Co., 122 N.E. 378, 381 (N.Y. 1919). The clause at issue
was a bar on oral modification. See id.
128. See, e.g., PATRICK BOLTON & MATHIAS DEWATRIPONT, CONTRACT THEORY 573–75 (2005);
Kevin E. Davis, The Demand for Immutable Contracts: Another Look at the Law and Economics of
Contract Modifications, 81 N.Y.U. L. REV. 487, 494–504 (2006) (describing four situations in which
immutability would be theoretically desirable); Christine Jolls, Contracts as Bilateral Commitments: A
New Perspective on Contract Modification, 26 J. LEGAL STUD. 203, 210–24 (1997); Alan Schwartz &
Robert E. Scott, Contract Theory and the Limits of Contract Law, 113 YALE L.J. 541, 611–14 (2003).
129. See Jolls, supra note 128, at 209–19. Cf. Levitin & Twomey, supra note 20, at 89 (suggesting
that servicers would be more likely to accept prohibitions of modification-barring provisions if such
prohibitions were accompanied by clarifications that servicers owed duties of care to all beneficiaries as
opposed to individual tranches).
the parties together.130

The demand for de facto immutability, however, remains in bond
contracts: creditors seek to deter opportunistic default (unwillingness, as
distinct from inability, to pay); debtors want to lower borrowing costs and
agree to forgo modification up front to signal willingness to pay;131 and
both parties discount heavily the possibility of insolvency, or inability to
pay, and any potential recovery values. By definition, neither party
accounts for the spillover effects of immutability, as discussed in Part V
below.
The literature on bond contract modification is steeped in history and
market particulars. The basic parameters of the bond debate were in place
half a century before the latest theoretical turn in favor of immutability.
Advocates of immutability in bond contracts sought to protect bondholder
minorities from corrupt, ignorant, and/or passive majorities.132 Arguing
against immutability, proponents of composition—or debt restructuring—
worried that minority holdout creditors might push an otherwise viable firm
into liquidation.133 This could result in spillovers and deadweight losses not
offset by anyone’s gains. The composition proponents therefore advocated
flexible contracts, including low-threshold amendment terms.134
A high-profile inflection in the debate occurred in the 1930s, when
Securities and Exchange Commission (“SEC”) Chairman (later–Supreme
Court Justice) William O. Douglas reacted to reports of corporate insiders
buying and voting bonds to appropriate for equity a recovery that rightly
belonged to debt.135 Douglas advocated what became the unanimity
130. See Davis, supra note 128 (noting coordination problems in contracts involving multiple
parties).
131. The signaling mechanism may be ineffective where, for example, good and bad borrowers
alike can easily adopt contracts that bar modification. See, e.g., Anna Gelpern & Mitu Gulati, Public
Symbol in Private Contract: A Case Study, 84 WASH. U. L. REV. 1627, 1712 (2006).
132. See Mark J. Roe, The Voting Prohibition in Bond Workouts, 97 YALE L.J. 232, 250–52
(1987) (noting that William O. Douglas and the Securities and Exchange Commission intended
immutability to induce bankruptcy because of their distrust of insider control of bond issues).
133. Id. at 235–39 (referring to minority holdups as the “buoying up effect”). See also Michael W.
McConnell & Randal C. Picker, When Cities Go Broke: A Conceptual Introduction to Municipal
Bankruptcy, 60 U. CHI. L. REV. 425, 449–50 (1993) (citing municipal bondholder complaints about
unanimity as an obstacle to compromise in 1933).
134. See, e.g., Roe, supra note 132, at 266–67 (discussing the flaws of immutability as a
“bondholder protection”).
135. See H. COMM. ON INTERSTATE & FOREIGN COMMERCE, H.R. REP. NO. 1016, TRUST
INDENTURE BILL OF 1939, at 24, 30–33 (1939); S. COMM. ON BANKING & CURRENCY, S. REP. NO. 248,
TRUST INDENTURE ACT OF 1939, at 4–8 (1939). Roe and others observe that these stories were in fact
unusual because the law of negotiable instruments at the time gave each bondholder veto power over
the terms of his contract as a condition of negotiability. See, e.g., Lee C. Buchheit & G. Mitu Gulati,
requirement of the TIA,136 a victory for immutability. As noted earlier, the

law requires each creditor’s consent to amend payment terms. It thus gives
creditors effective veto power over compositions. Against this background,
debtors and creditors should have a binary choice between securing each
bondholder’s consent and restructuring in bankruptcy under judicial
supervision. The existence of the bankruptcy alternative was crucial: the
compromise did not forbid modification altogether but sought to confine it
to the public, judicial realm. For Douglas, the need to control insider abuse
justified the costs of bankruptcy.137
Although the TIA applies to RMBS, the debates that led to its
enactment have limited relevance in the securitization context because
bond-issuing shells either cannot or are extremely unlikely to reorganize in
bankruptcy. Under the current regime, their distress must be resolved by
contract. Recent controversy surrounding amendment restrictions in
sovereign bond contracts offers an analogy.
Like securitization SPVs, sovereign states cannot go bankrupt;
however, like corporate debtors, they may require composition to survive.
Amendment clauses in sovereign bond contracts attracted policy and
Sovereign Bonds and the Collective Will, 51 EMORY L.J. 1317, 1332 (2002); Roe, supra note 132, at
256–57 (citing Enoch v. Brandon, 220 N.Y.S. 294, 296 (1927), and contemporaneous treatises).
136. Trust Indenture Act (TIA) of 1939, 15 U.S.C. §§ 77aaa–77bbbb (2006).
137. The consent provisions of the TIA also served another function: they protected bondholders
against conflicts of interest between bond trustees and bond issuers. The TIA’s restrictions responded to
the widespread chicanery in the real estate bond market in the 1920s and ’30s. The likes of “Straus
bonds,” “Greenebaum bonds,” and “Miller bonds” financed the construction of Manhattan’s most
famous art deco skyscrapers. See JAMES GRANT, MONEY OF THE MIND: BORROWING AND LENDING IN
AMERICA FROM THE CIVIL WAR TO MICHAEL MILKEN 162–69 (1992). These were single-asset real
estate bonds, issued against the earning power of a particular mortgaged building, often not yet
completed. Real estate bonds of the ’20s and ’30s featured all manner of self-dealing and conflicts of
interest, including having the underwriter or the underwriter’s affiliate serve as bond indenture trustee.
A “pet trustee” did not serve as an effective monitor of bond performance or advocate for the
bondholders. Such malfeasances were a major impetus for the TIA’s enactment. Indeed, the Act’s
original list of conflicts of interest was essentially a description of the real estate bond industry. See TIA
§ 310, 15 U.S.C. § 77jjj.
Single-asset real estate bonds were in some ways forerunners of securitization as they provided a
dedicated cashflow to investors based on real estate mortgage payments. They also were forerunners of
the CDO—a securitization vehicle whose assets (which might be actively managed) consist heavily of
interests in other securitizations. Thus, S. W. Straus & Co. was, by the mid-1920s, marketing “collateral
trust bonds,” which were described by a court as “a potpourri of indifferent subordinate mortgages
owned by the borrower and pledged as security, besides debentures of corporations owning real estate.”
GRANT, supra, at 163. Interestingly, it appears that the market began to respond to the problems of
single-asset real estate bonds before the TIA. For example, the single-asset real estate bonds at issue in
the famous case of Aladdin Hotel, which were issued in 1938, contained a contractual unanimous
consent provision. Aladdin Hotel Co. v. Bloom, 200 F.2d 627, 628–30 (8th Cir. 1953). The Aladdin still
stands in downtown Kansas City.
academic attention during the financial crises of the mid-1990s. Most

sovereign bonds were then issued under New York law and, although
exempt from the TIA, required unanimous creditor consent to amend
payment terms.138 Policymakers and academics widely expected the
unanimity requirements to delay composition and encourage holdouts; the
absence of the bankruptcy valve would only exacerbate coordination
problems.139
As with corporate bonds140 and RMBS, the argument against rigidity
in sovereign bond contracts emphasizes externalities. Absent an orderly
renegotiation framework, a debt overhang can depress investment and
growth in a country for years; sovereign default can trigger economic
collapse, bank failures, job losses, and contagion in financial markets.141
Because everyone understood the macroeconomic policy risks, it was
thought that debtors and creditors would hold up other governments and
organizations, such as the International Monetary Fund (“IMF”), for side
payments, potentially imposing costs on U.S. taxpayers, among others.142
The counterargument for rigidity in this context tracks contract theory.
It evokes information asymmetries and the challenge of disciplining
sovereign debtors: it is hard to tell “good” borrowers from “bad” ones.143
138. The reasons for this are disputed and, in any case, irrelevant here. Sovereign bonds are
specifically exempt from the TIA. See 15 U.S.C. § 77dd. Conventional wisdom holds that unanimity
provisions were mindlessly copied from corporate bond indentures, see Buchheit & Gulati, supra note
135, at 1331, or naively inserted in Brady bonds, which represented restructured loans, to deter
redefault, see James Hurlock & Troy Alexander, The Fire Next Time: The Dangers in the Next Debt
Crisis, INT’L FIN. L. REV., Mar. 1996, at 14, 14–15 (discussing the development of Brady bonds).
139. See, e.g., Anne Krueger, First Deputy Managing Dir., IMF, International Financial
Architecture for 2002: A New Approach to Sovereign Debt Restructuring, Address at the National
Economists’ Club Annual Members’ Dinner (Nov. 26, 2001), available at http://www.imf.org/external/
np/speeches/2001/112601.htm.
140. See Roe, supra note 132, at 232–35.
141. See generally BARRY EICHENGREEN ET AL., CRISIS? WHAT CRISIS? ORDERLY WORKOUTS
FOR SOVEREIGN DEBTORS (1995) (arguing for majority modification provisions in sovereign bond
contracts); Hurlock & Alexander, supra note 138 (arguing that new rigid sovereign bonds will be more
difficult to manage in distress than older sovereign loans).
142. See, e.g., Jeremy Bulow & Kenneth Rogoff, Multilateral Developing-Country Debt
Rescheduling Negotiations: A Bargaining-Theoretic Framework (IMF, Working Paper No. 88/35,
1988), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=884729 (referring to sovereign
debt restructuring as a tripartite negotiation with creditor country taxpayers); John B. Taylor, Under
Sec’y of the Treasury for Int’l Affairs, Sovereign Debt Restructuring: A U.S. Perspective, Remarks at
“Sovereign Debt Workouts: Hopes and Hazards?” Institute for International Economics Conference
(Apr. 2, 2002), available at http://www.treasury.gov/press/releases/po2056.htm (suggesting an official
subsidy for debtors and creditors to adopt collective action provisions in sovereign bonds).
143. See supra note 128 and accompanying text. For a discussion of how to deal with a defaulting
sovereign, see, for example, Arturo C. Porzecanski, From Rogue Creditors to Rogue Debtors:
Implications of Argentina’s Default, 6 CHI. J. INT’L L. 311, 326–32 (2005) (examining the specific case
Fiscal ability and political willingness to pay are notoriously hard to

disentangle.144 Suing foreign governments is of limited use because the
assets available to satisfy a judgment are either immune145 or in the home
country and out of creditors’ reach. And like corporate insiders,
government affiliates (such as agencies and state-owned enterprises) are
not shy about buying and voting government bonds to the detriment of
private bondholders.146
For reasons that had more to do with political economy than contract
theory, advocates of composition prevailed in 2003 when the sovereign
bond documentation standard shifted away from unanimity.147 Perhaps the
most remarkable aspect of this shift was that eliminating formal contractual
rigidity appeared to make little difference. After 2003, borrowers adopted
majority modification provisions without regard to credit quality, casting
doubt on the signaling value of modification terms.148 Meanwhile,
countries found ways to default and restructure with or without
unanimity.149 At least for now, it appears that formal barriers to amendment
in sovereign debt contracts did not make them immutable. It is difficult to
tell what, if any, extra cost such barriers might have imposed on
composition.
This experience highlights the difference between formal rigidity and
of Argentina in the 1990s).

144. To some extent all sovereign debtors must be good debtors since, as Walter Wriston
observes, countries never go bankrupt. Walter B. Wriston, Banking Against Disaster, N.Y. TIMES, Sept.
14, 1982, at A27. See also FEDERICO STURZENEGGER & JEROMIN ZETTELMEYER, DEBT DEFAULTS AND
LESSONS FROM A DECADE OF CRISES 38 (2006) (arguing that the effort to disentangle ability and
willingness to pay may be futile).
145. This is the case, for example, with embassies and military bases.
146. See CIBC Bank & Trust Co. v. Banco Central do Brasil, 886 F. Supp. 1105, 1107 (S.D.N.Y.
1995) (considering a case in which the creditors of Brazil sought to prevent a Brazilian government
instrumentality from voting Brazilian government debt); William W. Bratton & G. Mitu Gulati,
Sovereign Debt Reform and the Best Interest of Creditors, 57 VAND. L. REV. 1, 56–60 (2004) (citing
Patrick Bolton & David S. Scharfstein, Optimal Debt Structure and the Number of Creditors, 104 J.
POL. ECON. 1 (1996)); id. at 64–71 (arguing for intercreditor good faith duties); Buchheit & Gulati,
supra note 135, at 1339–42 (analyzing the facts of the CIBC case and considering the implications
regarding how courts will treat intercreditor duties of sovereign debtors). Some have argued that
“insider” status should extend to all institutions regulated by the sovereign borrower. Emerging Mkts.
Creditors Ass’n, Model Covenants for New Sovereign Debt Issues 1 (May 3, 2002) (on file with
authors).
147. See Gelpern & Gulati, supra note 131, at 1628–29.
148. See id. at 1712.
149. For example, Ecuador, Argentina, and Uruguay, among others, established corporate
restructuring tools, such as exchange offers and exit consents, to limit the impact of unanimity. See
STURZENEGGER & ZETTELMEYER, supra note 144, at 147–201, 211–26. See also Buchheit, supra note
123.
effective immutability in contracts. Some writers—notably Mark Roe and

Kevin Davis—make a point of distinguishing the two.150 Each stresses that
institutional design and covenants that stop short of banning modification
can go a long way toward immutability. Conversely, formal prohibition
might mean little where restructuring techniques and market norms let
parties circumvent their contracts—a view borne out by the sovereign bond
experience.
As noted in Part III, obstacles to RMBS modification are not just
formal; they are also structural and functional. Formal rigidity is a creature
of contract and statute. It can be direct, as in a term prohibiting
modification, or less so, as in the TIA’s voting thresholds and PSAs’
buyback requirement for modified mortgages. In either case, such
provisions impose formal constraints or conditions on renegotiation.
Structural rigidity is a product of institutional design. For example, the
trust, off–balance sheet accounting, and pass-through tax forms, as well as
the bankruptcy-remote features of RMBS, present obstacles to
modification, even where it is not their principal goal. Functional rigidity
goes to the economic incentives of contracting parties. RMBS features such
as tranching, resecuritization, and insurance are not designed primarily to
preclude modification, yet they create coordination problems and powerful
disincentives for junior creditors and insurers to cooperate in any
renegotiation.
RMBS recall sovereign bonds both in their bankruptcy remoteness,
and in some of the spillover effects from their enforcement in distress.151
Like corporations, securitization vehicles have multiple classes of creditors
that are vulnerable to holdup tactics and other collective action
problems.152 Trust organization, tax and accounting features,
senior/subordinate tranching, and resecuritization add more layers of
structural and functional rigidity on top of statutory and contractual barriers
to modification. This combination makes RMBS more effectively
immutable than either corporate or sovereign bonds. RMBS rigidity in the
face of real financial distress creates negative externalities, which is the
subject of Part V.
150. See Davis, supra note 128; Roe, supra note 132. Moreover, while Roe favors contractual
composition and majority voting, Davis appears to take no position on immutability as such, even as he
comes out against enforcement of no-amendment clauses.
151. See infra Part V.
152. See, e.g., Kurt Eggert, Comment on Michael A. Stegman et al.’s “Preventative Servicing Is
Good for Business and Affordable Homeownership Policy”: What Prevents Loan Modifications?, 18
HOUSING POL’Y DEBATE 279, 290–91 (2007) (describing how tranche warfare acts as a barrier to
modification).
B. RIGID CONTRACTS IN BANKRUPTCY THEORY
The quest for immutability in securitization design is consistent with

contract theory’s prediction that parties may commit to forgo ex post
modification in exchange for savings ex ante. It is also consistent with a
strain of bankruptcy theory, known as contractual bankruptcy,153 which
predicts that parties to business contracts will seek to opt out of the
Bankruptcy Code in favor of contractual ordering of the disposition of an
insolvent firm’s assets. Advocates of contractual bankruptcy argue that
such private ordering is more efficient than the public ordering mandated
by Congress.154 Private contracts can be drafted to reflect the idiosyncratic
preferences of particular debtors and creditors as to the relative values of
future bankruptcy protection and current cost of credit. The debate over the
relative merits of public and private ordering has dominated bankruptcy
scholarship for over a quarter century, beginning with Douglas Baird and
Thomas Jackson’s “creditors’ bargain” theory of bankruptcy.155
From the creditors’ bargain theory, several proposals for contractual
bankruptcy emerged. Early proposals did not advocate contracting out of
bankruptcy per se (which was disfavored by the courts156), but rather
proposed using bankruptcy as a mechanism for enforcing the
prebankruptcy, contractually determined priority scheme and for avoiding
rent extraction by managers and “out-of-the-money” claimants. Thus, Baird
envisioned a reformed Chapter 11 that would be used as a vehicle to
153. Elizabeth Warren & Jay Lawrence Westbrook, Contracting Out of Bankruptcy: An Empirical
Intervention, 118 HARV. L. REV. 1197, 1199 & n.6 (2005).
154. See, e.g., Robert K. Rasmussen, Debtor’s Choice: A Menu Approach to Corporate
Bankruptcy, 71 TEX. L. REV. 51, 53 (1992) (“[A] firm’s ability to file for bankruptcy reorganization
should be determined by the firm’s investors rather than by the government.”); Warren & Westbrook,
supra note 153, at 1201 (stating that contractualists believe “that a bankruptcy regime negotiated in the
marketplace will be far more efficient than the standardized ‘contract’ provided by Congress in the
Bankruptcy Code”).
155. The creditors’ bargain theory explains bankruptcy law as reflecting the bargain that creditors
would reach about the disposition of an insolvent firm absent coordination problems. See, e.g., Thomas
H. Jackson, Bankruptcy, Non-Bankruptcy Entitlements and the Creditors’ Bargain, 91 YALE L.J. 857,
860 (1982). The driving insight of the creditors’ bargain theory is that bankruptcy is designed to
overcome a common pool problem and prevent a destructive race to seize the firm’s assets. See, e.g.,
id.; JACKSON, supra note 67, at 7–19; Douglas G. Baird & Thomas H. Jackson, Bargaining After the
Fall and the Contours of the Absolute Priority Rule, 55 U. CHI. L. REV. 738 (1988) (explaining how the
absolute priority rule deals with these issues in various contexts); Douglas G. Baird & Thomas H.
Jackson, Corporate Reorganizations and the Treatment of Diverse Ownership Interests: A Comment on
Adequate Protection of Secured Creditors in Bankruptcy, 51 U. CHI. L. REV. 97, 105–09 (1984)
[hereinafter Baird & Jackson, Corporate Reorganizations]; Douglas G. Baird, Loss Distribution, Forum
Shopping, and Bankruptcy: A Reply to Warren, 54 U. CHI. L. REV. 815 (1987).
156. SCHWARCZ ET AL., supra note 22, at 54.
conduct going-concern sales of firms free and clear of claims, with the sale
proceeds then being divided among creditors according to absolute
priority.157 Variations on the sale method were proposed by Lucien Arye
Bebchuk;158 by Philippe Aghion, Oliver Hart, and John Moore;159 and by
Barry Adler.160 These proposals added increasing flexibility to the sale
process, but all envisioned a mandatory, judicially supervised process that
enforced a contractually determined, nonbankruptcy priority system.
Michael Bradley and Michael Rosenzweig took the idea of contractual
bankruptcy a step further and proposed removing it from the courts.161
They merely followed the logic of making Chapter 11 a mandatory sale
process. This process would allow creditors to bargain up front for where
they would fit in the firm’s priority structure, and the outcome would
reflect the idiosyncratic preferences of any particular creditor/debtor pair.
157. Douglas G. Baird, The Uneasy Case for Corporate Reorganizations, 15 J. LEGAL STUD. 127,
145–46 (1986).
158. See Lucian Arye Bebchuk, A New Approach to Corporate Reorganizations, 101 HARV. L.
REV. 775, 776–77 (1988). Bebchuk called for an automated bankruptcy, aimed at avoiding a fire sale
and a judicial valuation. Id. In Bebchuk’s design, absolute priority (determined ex ante by contract)
would be enforced in all reorganizations and liquidations through a system in which existing classes of
creditors and equity holders would be required to purchase all senior interests at face value or forfeit
their own interest. Id. at 785–88. The lowest priority class to bid in would own the company, with
everyone senior being paid in full. Id. at 782–83, 785–88. The seniority of classes of creditors would be
contractually determined and would include the bankruptcy bidding option. See id. at 781–88.
159. See Philippe Aghion, Oliver Hart & John Moore, The Economics of Bankruptcy Reform, 8 J.
L. ECON. & ORG. 523 (1992). Aghion, Hart, and Moore responded to Bebchuk’s proposal with one in
which a judge would allocate rights in the bankrupt firm among the claims holders and solicit bids for
control of the firm. Id. at 524. As they explained,
[The] proposed scheme is a decentralized variant on Chapter 7, in which noncash (as well as
cash) bids are allowed, and ownership of the firm is homogenized (to all equity), so that the
owners can decide (by vote) which of the bids to accept. However, insofar as noncash bids
allow for reorganization/recapitalization, [the] proposal can also be viewed as a decentralized
version of Chapter 11, in which conflicts of interest among different claimant groups are
avoided through homogenization of ownership.
Id. (footnote omitted).
160. See Barry E. Adler, Financial and Political Theories of American Corporate Bankruptcy, 45
STAN. L. REV. 311, 319–24 (1993) [hereinafter Adler, Financial Theories]. Adler rejects the common
pool justification for bankruptcy and proposes an alternative automated bankruptcy regime in which an
insolvent firm’s equity is transferred to the highest-priority class of creditors that cannot be paid on time
rather than using a bidding process. This process would be specified under the terms of financial
instruments issued by the firm. See Barry E. Adler, Finance’s Theoretical Divide and the Proper Role
of Insolvency Rules, 67 S. CAL. L. REV. 1107, 1110 & n.13 (1994) (arguing that “Chameleon Equity”—
a multipriority contractual hierarchy of preferred equity—would better resolve financial distress than
bankruptcy legislation); Adler, Financial Theories, supra, at 323–33 (same); Barry E. Adler, A Theory
of Corporate Insolvency, 72 N.Y.U. L. REV. 343, 352–57 (1997) (same); Barry E. Adler, A World
Without Debt, 72 WASH. U. L.Q. 811 (1994) (same).
161. See Michael Bradley & Michael Rosenzweig, The Untenable Case for Chapter 11, 101 YALE
L.J. 1043, 1078 (1992).
Robert Rasmussen162 and Alan Schwartz163 have each since proposed less
drastic contract bankruptcy alternatives that contemplate a substantial
degree of private autonomy for handling firm distress. While there are
significant differences among various contractual bankruptcy proposals,
they all require a degree of rigidity in their contractual design in order to
ensure that creditors actually get the benefit of their bargain. This aspect of
contractual bankruptcy is particularly stark in the proposals that stress
mandatory, automatic implementation and minimize the role of the courts.
Contractual bankruptcy has been criticized on the ground that it fails
to consider anything other than efficiency.164 Furthermore, its central
162. Rasmussen has proposed a different model of contract bankruptcy where firms may choose a
bankruptcy regime in their organizational charter from a menu of state-provided options. The choice
would be locked in unless every creditor consented to change it. Because creditors would know the
applicable bankruptcy regime before extending credit, they could price accordingly. The firm would
thus be able to balance between lower present costs of capital and greater bankruptcy protection in the
future, which would produce (absent hyperbolic discounting) the most efficient outcome. Involuntary
creditors would continue to be protected by mandatory rules, since they would not price credit based on
the organizational choices of their debtor. See Rasmussen, supra note 154.
163. Schwartz has proposed another model of contractual bankruptcy, where firms and their
creditors choose between a contractually locked-in (“renegotiation-proof”) insolvency regime and one
that is chosen by the firm (a “renegotiation contract” regime). Alan Schwartz, Essay, A Contract Theory
Approach to Business Bankruptcy, 107 YALE L.J. 1807, 1827–30 (1998) [hereinafter Schwartz,
Contract Theory Approach]. A “‘renegotiation-proof’ contract . . . will induce the firm to choose the
optimal bankruptcy system in the event of insolvency.” Id. at 1827. Absent the entry of a new creditor,
however, the terms of the contract to ensure the optimal bankruptcy regime would remain locked in.
“The contract is called renegotiation proof because no party will have an incentive to propose changes
in it in light of later events.” Id. Alternatively, in a “renegotiation contract” regime, creditors rely on the
unincentivized firm to choose the optimal bankruptcy regime. If reorganization is likely to have a
sufficiently higher value than liquidation, creditors would do better under a renegotiation regime. To
address the concern that the optimal bankruptcy regime may change over time, Schwartz introduces a
middle ground between lock-in and firm choice—a “partially renegotiation-proof contract” in which a
firm’s creditors’ renegotiation-proof contracts are adjusted to reflect the deal negotiated by the newest
creditor of the firm. See id. at 1831. Such a readjustment would ensure that the firm remains
incentivized to choose the optimal regime. See id. at 1827. To ensure an optimal renegotiation-proof
contract, the firm must be bribed to permit it to keep a percentage of the insolvency monetary return,
regardless of the specific regime. This will incentivize the firm to choose the regime that maximizes
monetary returns so long as it is a high enough percentage to offset any private benefits the firm wishes
to consume. See id. at 1827. See also Alan Schwartz, Essay, Bankruptcy Contracting Reviewed, 109
YALE L.J. 343, 343 (1999) (“If the rule against contracting for a preferred bankruptcy system were
relaxed, parties would write ‘bankruptcy contracts’ that would induce a borrowing firm to choose the
system that would be optimal for it and its creditors were it to become insolvent.”).
164. See, e.g., Christopher W. Frost, Bankruptcy Redistributive Policies and the Limits of the
Judicial Process, 74 N.C. L. REV. 75, 85–91 (1995) (describing the debate over the efficiency as a valid
criterion); Donald R. Korobkin, Rehabilitating Values: A Jurisprudence of Bankruptcy, 91 COLUM. L.
REV. 717, 762 (1991) (“At best, the economic account offers an undermining explanation of a
bankruptcy system that recognizes noneconomic outcomes as independent values. At worst, the
economic account does not explain ‘bankruptcy law’ at all, but merely restates its own economic
assumptions. . . . In contrast, the value-based account is founded on a deeper understanding of the
efficiency claim has been called into question. Some scholars have
questioned whether contractual bankruptcy would create new inefficiencies
from redistribution, offsetting any efficiency gains.165 Others have asked
whether the transaction costs involved in a contractual bankruptcy regime
would outweigh the efficiency gains.166 In a recent article, Elizabeth
Warren and Jay Lawrence Westbrook test the efficiency of contractual
bankruptcy in a multidistrict empirical study of business bankruptcy filings
from 1994.167 Warren and Westbrook find support for both the
redistribution and the transaction costs critiques of contractual bankruptcy;
however, their findings extrapolate from federal bankruptcy data168—as the
law stands, real firms may not contract out of bankruptcy.
Securitization offers a rare natural experiment for the contractual
approach.169 The bankruptcy-remote SPV (the firm) functionally opts out
concern to which bankruptcy law is addressed. Bankruptcy law is a response to the problem of financial
distress—not only as an economic, but as a moral, political, personal, and social problem that affects its
participants.”); Donald R. Korobkin, The Role of Normative Theory in Bankruptcy Debates, 82 IOWA L.
REV. 75 (1996) (examining the debate over normative justifications for bankruptcy policy); Lawrence
Ponoroff & F. Stephen Knippenberg, The Implied Good Faith Filing Requirement: Sentinel of an
Evolving Bankruptcy Policy, 85 NW. U. L. REV. 919, 962 (1991) (counseling against “any closed-end
theory or understanding of the law”); Elizabeth Warren, Essay, Bankruptcy Policymaking in an
Imperfect World, 92 MICH. L. REV. 336, 336–40 (1993) [hereinafter Warren, Bankruptcy
Policymaking]. Other important considerations have been proposed. See, e.g., Jean Braucher,
Bankruptcy Reorganization and Economic Development, 23 CAP. U. L. REV. 499, 517–18 (1994)
(arguing for the need to protect employee interest because of inability to diversify labor capital); Donald
R. Korobkin, Employee Interests in Bankruptcy, 4 AM. BANKR. INST. L. REV. 5 (1996) (arguing for the
need to protect employee interest because of inability to diversify labor capital); Raymond T. Nimmer,
Negotiated Bankruptcy Reorganization Plans: Absolute Priority and New Value Contributions, 36
EMORY L.J. 1009, 1032–34 (1987); Warren, Bankruptcy Policymaking, supra, at 352–61 (suggesting
various distributional goals besides economic efficiency and emphasizing the need to internalize costs
of business failure); Elizabeth Warren, A Theory of Absolute Priority, 1991 ANN. SURV. AM. L. 9
(arguing for the need to account for the interests of the community, including potential customers,
suppliers of the firm, and taxing authorities).
165. See, e.g., Lucian Arye Bebchuck & Jesse M. Fried, The Uneasy Case for the Priority of
Secured Claims in Bankruptcy, 105 YALE L.J. 857 (1996); Susan Block-Lieb, The Logic and Limits of
Contract Bankruptcy, 2001 U. ILL. L. REV. 503, 548–49 (citing transaction costs); Lynn M. LoPucki,
The Case for Cooperative Territoriality in International Bankruptcy, 98 MICH. L. REV. 2216, 2243
(2000); Jay Lawrence Westbrook, The Control of Wealth in Bankruptcy, 82 TEX. L. REV. 795, 830–37
(2004). See generally Lynn M. LoPucki, Essay, Contract Bankruptcy: A Reply to Alan Schwartz, 109
YALE L.J. 317 (1999) (criticizing Schwartz’s model); Warren & Westbrook, supra note 153 (using
empirical data to critique the contractualist position).
166. See Rasmussen, supra note 154, at 114–16 (discussing the adverse selection problem);
Warren & Westbrook, supra note 153, at 1201–02.
167. Warren & Westbrook, supra note 153, at 1201–02.
168. See id. at 1202.
169. See Randal C. Picker, Security Interests, Misbehavior, and Common Pools, 59 U. CHI. L.
REV. 645, 649–53 (1992) (noting various ways in which secured credit can serve as a bankruptcy
contract); Steven L. Schwarcz, Rethinking Freedom of Contract: A Bankruptcy Paradigm, 77 TEX. L.
of federal bankruptcy law. The rights of the creditors of the SPV are
governed by the PSA, in particular its tranching and payment-schedule
provisions. The senior/subordinate structure of tranching creates an
absolute priority regime for the creditors of the SPV. Creditors are free to
bargain for the terms of the securities that they buy—different SPVs have
different tranching arrangements—and investors are free to purchase the
position they wish to have in the SPV’s capital structure. Securitization
even deals with the problem of involuntary creditors who cannot “bargain”:
because the SPV is essentially a passive entity that merely holds assets, it is
unlikely to incur liability to involuntary creditors.170 In sum, securitization
effectively creates a self-executing contractual bankruptcy regime; this vast
market shows that Schwartz was correct when he observed that
“bankruptcy contracting would occur if it were permitted.”171
Securitization’s natural experiment in contractual bankruptcy appears
to be successful. Firms will securitize only if securitization offers a lower
cost of capital than other methods of finance. Securitization will only offer
a lower cost of capital than direct debt or equity financing if it offers
creditors greater benefits. The very existence of securitization proves its
efficiency, at least from the perspective of the firm. Of course, this
efficiency is due to more factors than bankruptcy remoteness and the
certainty of priority in distribution; however, as noted earlier, these two
factors are central to securitization’s design and are perhaps the most
important features of securitization for investors.172
Yet the experience with mortgage securitization also gives reason for
skepticism about the efficiency claims of contractual bankruptcy. The
mortgage crisis has revealed a third category of problems with contractual
bankruptcy in addition to redistribution and transaction costs: this category
comprises spillover effects from locking the parties into a precomitted
resolution framework, or contract rigidity. Because contractual bankruptcy
models limit themselves to the context of the firm, they do not consider
spillovers beyond the firm.173 Contractualists leave the problem to the
REV. 515, 597–99 (1999) (suggesting that securitization transactions should be enforceable as
bankruptcy waiver contracts). Cf. Schwartz, Contract Theory Approach, supra note 163, at 1833
(“Because bankruptcy contracts are currently illegal, there is no data about real contracts that could
support [the argument that differences in creditor preferences of bankruptcy systems could be
overcome].”).
170. This is sometimes reinforced by covenants. See supra note 75. The major exception would be
claims and counterclaims from obligors on the SPV’s assets.
171. Schwartz, Contract Theory Approach, supra note 163, at 1833.
172. See supra Part III.C–D.
173. See, e.g., Schwartz, Contract Theory Approach, supra note 163, at 1814.
market.174 But thick markets are not always there to pick up the slack, and
maximizing the value of a bankruptcy estate maximizes social welfare if
and only if the marginal increase in welfare from the particular bankruptcy
regime offsets the marginal decrease in welfare from the regime’s
externalities.175
The debates surrounding contractual bankruptcy have all been in the
context of business bankruptcy. There is an unspoken recognition that
consumer bankruptcy policy is different. The transaction costs of
contractual bankruptcy for consumers would be prohibitively high and
would necessitate accounting for numerous involuntary creditors, as well as
for creditors who cannot easily adjust their behavior for the many bespoke
bankruptcy regimes that would be available to consumers. The efficiency
benefits of a standardized bankruptcy regime for consumers are so manifest
that there has been no attempt to apply contractual bankruptcy to them.
And yet that is precisely what securitization accomplishes.
Securitization transforms consumer debt into business debt. The resulting
transformation has profound implications for the application of a
contractual bankruptcy regime, embodied in RMBS PSAs, designed to
inhibit the reworking of the securitized consumer debts, forcing consumer
debt out of the social policy world of the Bankruptcy Code and into the
private ordering of securitization contracts.176
In sum, securitization is the embodiment of contractual bankruptcy.
For the device to work, its contracts must be rigid. Contractual bankruptcy
theories, however, do not consider the full range of spillover effects created
by rigidity. The problem of spillover effects is particularly salient in
residential mortgage securitization for two reasons. First, the enormous size
of the market creates the risk that any disruptions would affect the broader
financial system and the macroeconomy. Second, because residential
174. E.g., id. at 1817 (“It is unnecessary for bankruptcy law to protect communities when thick
markets exist. In a thick market, there are good substitutes for the firm’s performance.”); Baird &
Jackson, Corporate Reorganizations, supra note 155, at 101–02 (“The failure of a firm affects many
who do not, under current law, have cognizable ownership interests in the firm outside of bankruptcy.
The economy of an entire town can be disrupted when a large factory closes. Many employees may be
put out of work. The failure of one firm may lead to the failure of those who supplied it with raw
materials and those who acquired its finished products. Some believe that preventing such
consequences is worth the costs of trying to keep the firm running and justifies placing burdens on a
firm’s secured creditors.” (footnote omitted)).
175. So thoroughly has the firm-centered framing of the bankruptcy debate taken root that even its
most vigorous critics, such as Warren, ultimately succumb to it after expanding the list of constituents
slightly to include communities surrounding the firm. See, e.g., Warren, Bankruptcy Policymaking,
supra note 164, at 352–61 (suggesting various distributional goals besides economic efficiency).
176. See supra notes 104–17 and accompanying text.
mortgage securitization effectively transforms consumer debt into business

debt, it imparts elements of contractual bankruptcy to relationships that it
was never meant to cover. The next part elaborates on spillover effects
from securitization.
V. SPILLOVERS: CONTRACTS AS SUICIDE PACTS
As discussed in Part III, RMBS PSAs constrain the modification of

two kinds of contracts. First, they restrict amendment of underlying
mortgage loans. Second, they restrict amendment of the RMBS themselves.
Combined, these restrictions create four sets of negative spillover effects:
for communities, for other creditors, for the financial markets, and for the
economy as a whole. These spillover effects stem from pervasive
foreclosures, from complexity and illiquidity in RMBS, and from
downward pressure on asset prices. PSA design also makes it difficult for
the government to regulate securitization contracts to mitigate the negative
spillovers.
Before proceeding to catalogue PSAs’ effects on the economy and
society, it is useful to set out their effect on the debtor-homeowner, who is
not party to the PSA contract and does not know its terms. A homeowner
whose loan has been securitized under a restrictive PSA does not have the
same workout options as one whose loan has not been securitized, or one
whose loan is sold under a PSA that is more amenable to modification.177 If
this homeowner cannot pay his or her loan according to its original terms,
the homeowner is more likely to lose the house. Whether homeowners do,
would, or could get a price break to compensate them for this commitment
is an open question. Homeowners and lenders do not know at the time of
origination whether the loan will be securitized, much less under what
terms. But even if they had perfect information, borrowers might not ask
for or receive proper compensation.178 Where securitization under
restrictive PSAs is pervasive, it is unlikely that contractual rigidity would
help creditors detect bad borrowers ex ante: bad apples would not opt out.
More importantly for our purposes, even if homeowners asked for and
received a better interest rate in exchange for giving up ex post workout
177. Compare Adelino et al., supra note 125, at 14, 25–26 (finding a small role for contract
frictions in the context of renegotiation and suggesting that modifications were not necessarily in the
best interest of investors), with Cong. Oversight Panel, supra note 125 (noting differences in
modification type based on securitization status).
178. See Levitin & Twomey, supra note 20, at 73–75 (discussing cognitive issues in mortgage
lending). See also Bar-Gill, supra note 2 (explaining how the contractual design features of subprime
loans can be explained as responses to the imperfect rationality of borrowers).
options, their arrangement might be costly for others who have nothing to
do with either the mortgage or the securitization contracts. The remainder
of this part considers the effects of residential mortgage securitization on
the broader economy and society.
First, where most home loans are securitized under restrictive PSAs,
one would expect more foreclosures in an economic downturn. A high
foreclosure rate creates negative externalities for communities.
Foreclosures impose new costs on communities, as foreclosed properties
are often magnets for crime and fire;179 push down the price of neighboring
properties;180 and reduce property tax revenue for local governments.181
Foreclosures also have unquantifiable costs as debtors’ relocation affects
social relationships.182 Foreclosures have even been linked to public health
problems such as the spread of the West Nile Virus.183
Second, restrictive PSAs can contribute to a collective action problem
of the sort that produces bank runs.184 Where there is a wave of
foreclosures, the real estate market becomes flooded with properties,
pushing down home prices. Creditors that might not have foreclosed
otherwise, rush to salvage what is left of their investment. Mass
foreclosures can create a negative feedback loop that begets more
foreclosures and greater losses for lenders. Thus, restrictive PSAs impose
costs on RMBS holders and on mortgagees generally, quite apart from the
179. See, e.g., WILLIAM C. APGAR & MARK DUDA, COLLATERAL DAMAGE: THE MUNICIPAL
IMPACT OF TODAY’S MORTGAGE FORECLOSURE BOOM 6 (2005), available at http://www.995hope.org/
content/pdf/Apgar_Duda_Study_Short_Version.pdf; Dan Immergluck & Geoff Smith, The Impact of
Single-Family Mortgage Foreclosures on Neighborhood Crime, 21 HOUSING STUD. 851, 855–56
(2006).
180. See, e.g., APGAR & DUDA, supra note 179, at 5; DAN IMMERGLUCK & GEOFF SMITH,
WOODSTOCK INST., THERE GOES THE NEIGHBORHOOD: THE EFFECT OF SINGLE-FAMILY MORTGAGE
FORECLOSURES ON PROPERTY VALUES (2005), available at http://www.woodstockinst.org/
index.php?option=com_docman&task=doc_download&gid=52 (estimating that in Chicago the 3750
foreclosures that took place between 1997 and 1998 reduced surrounding property values by almost $6
million); Dan Immergluck & Geoff Smith, The External Costs of Foreclosure: The Impact of Single-
Family Mortgage Foreclosures on Property Values, 17 HOUSING POL’Y DEBATE 57 (2006) (same).
181. See GLOBAL INSIGHT, THE MORTGAGE CRISIS: ECONOMIC AND FISCAL IMPLICATIONS FOR
METRO AREAS 2 (2007), available at http://www.vacantproperties.org/resources/documents/
USCMmortgagereport.pdf; John Kroll, Foreclosure Study Says Vacant Properties Cost Cleveland $35+
Million, PLAIN DEALER (Cleveland), Feb. 19, 2008, available at http://blog.cleveland.com/metro/2008/
02/foreclosure_study_says_vacant.html.
182. See Adam J. Levitin, Helping Homeowners: Modification of Mortgages in Bankruptcy, 3
HARV. L. & POL’Y REV. ONLINE 1, 1 (2009), http://www.hlpronline.com/Levitin_HLPR_011909.pdf;
Levitin, supra note 46, at 569.
183. See Daniel Denoon, Foreclosures Worsen Spread of West Nile, CBSNEWS.COM, Oct. 23,
2008, http://www.cbsnews.com/stories/2008/10/02/health/webmd/main4495947.shtml.
184. See Garrett Hardin, The Tragedy of the Commons, 162 SCIENCE 1243 (1968).
effects on the holders and mortgagees under any given PSA.

Third, restrictive PSAs can help drive down the financial markets.
U.S. RMBS are among the world’s most widely held securities; trillions of
dollars in derivative products further amplify and transmit their effects.
Despite its size, the RMBS market in its present form is very young
compared to the corporate bond market. Its pricing models rely on a
relatively short performance history and a very thin market
infrastructure.185 Comparing the RMBS and corporate bond markets’
handling of failure is instructive. Although the RMBS pricing models take
into account the possibility of nonpayment on an occasional mortgage, they
do not account for large-scale failure. As noted in Part IV.B, the securities
are designed to fit models where failure is near impossible, or at worst,
precisely compartmentalized. These models are ill equipped to predict
recovery values from widespread foreclosures. Moreover, servicers have
never gone through a foreclosure epidemic in a downturn; they lack the
administrative capacity to process foreclosures on a large scale. In contrast,
corporate bonds benefit from a long performance history and an established
infrastructure to handle distress, including numerous renegotiation options.
In the worst case, creditors can look to the company’s liquidation value. It
would take a combination of economic collapse and a breakdown of the
U.S. bankruptcy system for corporate liquidation values to lose meaning. In
contrast, even in good economic times, private-label RMBS rarely trade;
they are illiquid, noncommodity products. A relatively moderate downturn
might be expected to render RMBS and the structured products based on
RMBS even harder to value.186
Securities whose recovery value is difficult or impossible to determine
either trade at a deep discount or become illiquid. Illiquidity in the RMBS
market reverberates worldwide. As financial institutions incur losses from
exposure to RMBS and market risk, they may respond by hoarding cash,
which in turn squeezes consumer and business credit. A downward spiral
ensues.
Fourth, brittle PSAs also have implications for the economy as a
whole. This category of spillover effects is the macroeconomic counterpart
of two categories already described: foreclosures on a large scale depress
185. See, e.g., Gillian Tett, Aline Van Duyn & Paul J. Davies, A Re-Emerging Market?: Bankers
Are Seeking Simpler Ways to Sell on Debt, FIN. TIMES (London), July 1, 2008, at 9 (noting the decades-
old emergence of securitization and its significant growth since 2000).
186. This discussion does not address RMBS attributes that may make valuation difficult but that
do not relate to PSA rigidity. These include leverage, opacity, and poor due diligence. See, e.g., id.
aggregate spending power and real estate asset prices, and illiquidity in the
RMBS market fuels a downward price spiral in financial assets. Where real
estate prices are predicated on a liquid market in asset-backed securities,
they collapse with the securitization market. A credit crunch depresses
consumption and investment and, ultimately, growth.
Although securitization contracts generate significant externalities and
impose costs on a wide range of constituencies beyond the contracting
parties, they are designed to limit the government’s capacity to mitigate
their potential adverse impact on the economy. Bankruptcy remoteness,
tranching, and resecuritization limit intervention options in distress. Of
these, bankruptcy remoteness is the most obvious constraint on regulation
and crisis response since by its terms this feature eliminates the state’s role
in managing financial distress. Securitization replaces a statutory
bankruptcy regime with a contractual one that is more brittle, less
transparent, and largely immune to social policy considerations.187
The next part reviews historical experience in responding to contracts
with significant spillover effects that warranted government intervention.
VI. OVERCOMING RIGIDITIES: NEW DEAL TOOLS
History reveals a menu of standard responses to rigid contracts. First,

there is statutory bankruptcy, designed and mandated by the legislature.
Under the Bankruptcy Code, the debtor can avoid antimodification features
in its own contracts.188 This is the method used to avoid contractual
rigidities in corporate bonds and farm mortgages.189 A similar approach
187. Other implications for government intervention are less visible. When the link between
debtor and creditor has been severed—replaced with impermeable, hyperrigid layers of securitization—
regulatory tools premised on the existence of that link become worthless. For example, regulatory
accounting treatment of a portfolio loan on the creditor’s books can have a direct effect on that
creditor’s willingness to restructure the loan. Marking assets to market creates an incentive to modify a
problem asset where renegotiation can produce a mark-to-market gain; if a bank is carrying an impaired
mortgage at 50 cents on the dollar, it has the incentive to grant a homeowner relief so long as the
modified instrument can be valued at more than 50 cents. Regulatory forbearance (for example, letting
the bank carry the loan at 80 cents) might force some loss-sharing with the borrower. The same
dynamic holds with conventional debt securities. In contrast, where assets are RMBS, changing the
accounting treatment on the investor’s books would have little or no impact on the prospects of
restructuring the underlying mortgage since the relationship between the value of the mortgage and the
value of the RMBS slice held by any given investor is indirect at best. (This is true even without taking
into account the formidable challenge of valuing RMBS in a depressed market.)
188. 11 U.S.C. § 365(a) (2006).
189. See JACKSON, supra note 67; infra Part VI.C.
was mooted to overcome rigidity by treaty in sovereign debt contracts.190

Second, governments can offer parties special incentives to
circumvent or change their contracts. These incentives can take the form of
sticks or carrots. Special incentives were used to deal with farm mortgage
foreclosures during the Great Depression: the government used a
combination of state foreclosure moratoria (sticks)191 and federal
refinancing subsidies (carrots). Again, analogues were considered for
sovereign debt contracts.192 Recent proposals directed at RMBS PSAs
follow suit. One such proposal would take away favorable tax treatment for
securitizations unless antimodification provisions were removed (a
stick).193 Others, including the Obama administration’s Making Home
Affordable Program, offer servicers bounties for every loan they modify
(carrots).194
190. An attempt to establish a sovereign bankruptcy regime by treaty using antirigidity

arguments—among others—failed in 2003. Statutory sovereign bankruptcy was a political nonstarter.
No state would cede authority over its debt management to an international body; no debtor or creditor
was prepared to accept an IMF-run regime (even a weak one); and no other governance options were on
the table. See Krueger, supra note 139 (presenting the IMF’s proposal of such a system in 2002). See
generally Brad Setser, IPD Task Force on Sovereign Debt, The Political Economy of the SDRM (June
8, 2009) (unpublished manuscript, on file with authors) (describing the political forces that led to the
failure of the IMF’s proposed sovereign debt restructuring mechanism (“SDRM”) on worldwide policy
debates). For normative objections to the sovereign bankruptcy regime, see Daniel K. Tarullo, Rules,
Discretion, and Authority in International Financial Reform, 4 J. INT’L ECON. LAW 613, 627–40
(2001). See generally Kenneth Rogoff & Jeromin Zettelmeyer, Bankruptcy Procedures for Sovereigns:
A History of Ideas, 1976–2001 (IMF, Working Paper No. 02/133, 2002) (on file with authors)
(surveying the recent history of sovereign bankruptcy proposals).
191. See Home Bldg. & Loan Ass’n v. Blaisdell, 290 U.S. 398, 415–16, 422–23 (1934).
192. European and Canadian governments advocated mandatory sovereign debt standstills. See
PAUL BLUSTEIN, THE CHASTENING: INSIDE THE CRISIS THAT ROCKED THE GLOBAL FINANCIAL
SYSTEM AND HUMBLED THE IMF 170–74 (2001); Andy Haldane & Mark Kruger, The Resolution of
International Financial Crises: Private Finance and Public Funds 10–15 (Nov. 2001) (unpublished
manuscript), available at http://www.bankofengland.co.uk/publications/other/financialstability/
boeandboc.pdf. U.S. officials raised the idea of paying the parties to sovereign bond contracts to adopt
majority modification provisions. See Taylor, supra note 142. On the other hand, the latest standard
U.S. bilateral investment treaties (“BITs”) deny expropriation protection to creditors under sovereign
bonds that require unanimous consent to amend payment terms. See, e.g., Treaty Concerning the
Encouragement and Reciprocal Protection of Investment, U.S.-Uru., Annex G, ¶ 2(b)(iii), Nov. 4, 2005.
193. See Michael S. Barr & James A. Feldman, Ctr. for Am. Progress, Issue Brief: Overcoming
Legal Barriers to the Bulk Sale of At-Risk Mortgages 2–3 (Apr. 2008), available at
http://www.americanprogress.org/issues/2008/04/pdf/reimc_brief.pdf.
194. See, e.g., U.S. Dep’t of Treasury, Homeowner Affordability and Stability Plan Fact Sheet 5
(Feb. 18, 2009), available at http://www.treas.gov/press/releases/20092181117388144.htm; Press
Release, U.S. Dep’t of Treasury, Neel Kashkari Remarks on GSE, HOPE Now Streamlined Loan
Modification Program (Nov. 11, 2008), available at http://www.treas.gov/press/releases/archives/
200811.html. See also Christopher Mayer, Edward Morrison & Thomasz Piskorski, A New Proposal for
Loan Modifications, 26 YALE J. ON REG. 417, 420 (2009).
Third, government can simply use its eminent domain power to seize
the contractual rights and slice through the Gordian knot of contractual
rigidities. Thus, Howell Jackson and Lauren Willis have both proposed
mass nationalization of mortgage loans from securitization vehicles, which
the government could then renegotiate without regard to contractual
limitations.195 The threat of nationalization, of course, can itself be used as
a stick to encourage voluntary renegotiations.
There are other ways of dealing with contractual rigidities that may
cause social harm. One approach involves narrowly tailored legislation that
renders the offending contractual language unenforceable on public policy
grounds. This was used in the New Deal to overcome gold indexation when
the United States devalued the dollar. Another approach, used to break up
utility holding companies beginning in 1935, involves broad and flexible
administrative mandates to simplify complex financial structures. In
addition, a combination of foreclosure moratoria and statutory bankruptcy
was used twice over the past century to help restructure troubled farm
mortgages that suffered from significant creditor coordination problems.196
These examples hold lessons for different aspects of PSA rigidity:
formal contractual restraint, organizational structure (notably, complexity),
and functional rigidity in the form of coordination problems. We
summarize the experience below.
A. GOLD CLAUSES AND FORMAL RIGIDITY
When Franklin Roosevelt took office on March 4, 1933, state-

mandated bank closures were spreading across the country.197 Thousands
of financial institutions failed or teetered on the brink. Where banks were
open, people lined up “with satchels and paper bags to take gold and
currency away from the banks to store in mattresses and old shoe
195. See Howell E. Jackson, Op-Ed, The Paulson Plan Should Target Bad Loans, Not Burned
Investors, CHRISTIAN SCI. MONITOR, Sept. 25, 2008, available at http://www.csmonitor.com/2008/
0925/p09s02-coop.html; Lauren E. Willis, Stabilize Home Mortgage Borrowers, and the Financial
System Will Follow 1–2 (Loyola Law Sch. L.A., Legal Studies Paper No. 2008-28, 2008), available at
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1273268. Both Jackson and Willis contend that the
federal government should exercise eminent domain, but eminent domain power could also be
exercised by the states. The potential expense of a widescale exercise of eminent domain is one factor
that reduces its appeal as a solution.
196. See infra Part VI.C.
197. BARRY EICHENGREEN, GOLDEN FETTERS: THE GOLD STANDARD AND THE GREAT
DEPRESSION, 1919–1939, at 329 (1992); WILLIAM E. LEUCHTENBURG, FRANKLIN D. ROOSEVELT AND
THE NEW DEAL 38–39, 42–43 (1963).
boxes.”198 Withdrawals accelerated on rumors of dollar devaluation,199

even as the president-elect remained coy about his monetary intentions.200
On March 5, 1933, President Roosevelt invoked the Trading with the
Enemy Act of 1917 to declare a national bank holiday and bar all
transactions in gold.201 A law to this effect followed days later.202 By mid-
April, Roosevelt announced that he would take the United States off the
gold standard.203 A new monetary framework passed within weeks as part
of a farm bill.204 It gave the executive discretion to inflate by remonetizing
silver, printing money, or changing the gold content of the dollar, but did
not mandate devaluation.205 As the year wore on, fears of “marching
farmers” and “an agrarian revolution” in New Deal policy circles eclipsed
the calls for stable money.206 On January 30, 1934, Congress enacted the
Gold Reserve Act, requiring a 40 percent minimum reduction in the value
of the dollar, and directing all gold coin to be melted into bullion.207
Roosevelt formally devalued the next day.208
From the start, the president’s monetary activism faced an obstacle in
private contracts. Roughly $100 billion in long-term bond contracts
contained what the New York Times described at the time as “the familiar
clause, ‘principal and interest payable in the United States gold coin of
present standard of weight and fineness.’”209 The formulation was
ubiquitous in obligations of the United States, foreign and subnational
198. LEUCHTENBURG, supra note 197, at 39.

199. EICHENGREEN, supra note 197, at 324–29. Roosevelt’s Attorney General Homer Cummings
later claimed that between February and early March, more than $476 million in gold had been
withdrawn from Federal Reserve banks and the U.S. Treasury, of which $311 million appeared to head
abroad. Oral Argument of Attorney General Homer Cummings at 265, Norman v. Balt. & Ohio R.R.,
294 U.S. 240 (1935) [hereinafter Cummings Argument].
200. See LEUCHTENBURG, supra note 197, at 38. See also Kenneth W. Dam, From the Gold
Clause Cases to the Gold Commission: A Half Century of American Monetary Law, 50 U. CHI. L. REV.
504, 504 (1983) (citing Roosevelt’s campaign criticism of Hoover’s alleged soft money leanings).
201. See Michal R. Belknap, The New Deal and the Emergency Powers Doctrine, 62 TEX. L. REV.
67, 73 (1983).
202. Emergency Banking Relief Act, ch. 1, §§ 2, 3, 48 Stat. 1, 1–2 (1933).
204. See Agricultural Adjustment Act, Pub. L. No. 73-10, § 43(b), 48 Stat. 31, 52–53 (1933).
205. Id. §§ 2, 43(b), 48 Stat. at 32, 52–53.
206. See ARTHUR M. SCHLESINGER, JR., THE COMING OF THE NEW DEAL, 237, 242 (First Mariner
Books 2003) (1958). See also LEUCHTENBURG, supra note 197, at 51.
207. Gold Reserve Act of 1934, ch. 6, Pub. L. No. 73-87, §§ 5, 12, 48 Stat. 337, 340, 342–43.
208. Proclamation No. 2072, 48 Stat. 1730 (Jan. 31, 1934).
209. Ignore Indenture “Payable in Gold”: Agents for Bonds with Coupons Due Fail to Give Coin
When Demand Is Made; Court Action Possible, N.Y. TIMES, May 2, 1933, at 2; Randall S. Kroszner, Is
It Better to Forgive Than to Receive? An Empirical Analysis of the Impact of Debt Repudiation 2 (Nov.
2003) (unpublished manuscript, on file with the University of Chicago Graduate School of Business).
governments, railroads, utilities, and corporations.210 Available gold supply

in the United States was reported at about $4 billion at the time; it was at
$11 billion worldwide—a small fraction of the amount ostensibly owing to
the creditors, if the clauses were to be read literally as promises to deliver
gold.211 The total amount of gold clause debt outstanding also far exceeded
the size of the U.S. economy.212 Enforcement of the term (as either gold or
gold value) in conjunction with dollar devaluation would have increased a
significant portion of the country’s public and private debt by nearly 70
percent and caused “mass bankruptcy.”213
The gold clauses represented a simple indexation mechanism to
protect creditors from devaluation, commonplace throughout history and
still popular in many parts of the world.214 In the United States, they gained
popularity in the wake of the monetary chaos of the Civil War.215 The gold
term itself was not a source of legal rigidity in the contemporary contract
theory sense: creditors could agree to change or abandon indexation.216 But
like many indexation devices, these clauses created economic rigidity: they
purported to lock the debtor into a commitment to pay a prespecified value
210. See Gold Obligations Are $100,000,000,000; Federal Bonds Total $22,000,000,000, N.Y.
TIMES, May 27, 1933, at 2.
211. Cummings Argument, supra note 199, at 255–56, 265. Whether the clauses were to be read
as promising payment in gold coin, or in paper dollars but in the amount equivalent to the gold value of
the debt at the time of the contract, was not entirely clear and was a subject of dispute in subsequent
litigation. See, e.g., id.; Henry M. Hart, Jr., The Gold Clause in United States Bonds, 48 HARV. L. REV.
1057, 1071 (1935).
212. See U.S. Dep’t of Commerce, Bureau of Econ. Analysis, National Income and Products
Accounts, Table 1.1.5: Gross Domestic Product, available at http://bea.gov/national/nipaweb/
SelectTable.asp (select Table 1.1.5) (listing U.S. GDP as $56.4 billion in 1933 and $66.0 billion in
1934).
213. Kroszner, supra note 209, at 2. See also Cummings Argument, supra note 199, at 256
(warning of a return to chaos).
214. For example, news reports in 1933 cite a British case construing similar clauses against the
creditor. See, e.g., Turner Catledge, Gold-Bond Clause Awaits Court Test, N.Y. TIMES, May 7, 1933, at
XX2 (reporting a case in which a British court substituted currency as an example of Britain’s
experience with gold clause debt). Creditors’ briefs in the subsequent U.S. Gold Clause Cases refer to
indexed Serbian and Brazilian debt. See, e.g., Oral Argument of Reconstruction Finance Corp. at 277,
Norman v. Balt. & Ohio R.R., 294 U.S. 240 (1935). The Court itself highlights German reparation
obligations. See Norman, 294 U.S. at 299 n.3. See also Arthur Nussbaum, Comparative and
International Aspects of American Gold Clause Abrogation, 44 YALE L.J. 53, 60–61 (1934) (noting the
numerous other countries that abrogated gold clauses). See generally KEITH S. ROSENN, LAW AND
INFLATION 130–54, 267–88 (1982) (discussing more recent domestic and international experiences with
contract indexation); Artur Nussbaum, Multiple Currency and Index Clauses, 84 U. PA. L. REV. 569,
579–582 (1936) (comparing the U.S. experience with indexation to that of European countries).
215. See Dam, supra note 200, at 507; Daniel W. Levy, A Legal History of Irrational Exuberance,
48 CASE W. RES. L. REV. 799, 856 (1998).
216. However, the prevailing view of the law of negotiable instruments at the time would have
made renegotiation difficult or impossible. See supra note 135.
notwithstanding inflation. And the ubiquity of the clause worked as a

policy constraint on the government. Congressman Steagall articulated the
spillover argument as follows in a House committee report, later quoted by
the Supreme Court:
These gold clauses render ineffective the power of the government to
create a currency. . . . If the gold clause applied to a very limited number
of contracts and security issues, it would be a matter of no particular
consequence, but in this country virtually all obligations, almost as a
matter of routine, contain the gold clause. . . . [N]o currency
system . . . can meet the requirements of a situation in which many
billions of dollars of securities are expressed in a particular form of the
circulating medium, particularly when it is the medium upon which the
entire credit and currency structure rests.217
Congress responded on June 5, 1933, with a joint resolution that rendered
the gold clauses unenforceable and allowed nominal payments “dollar for
dollar” to discharge the underlying obligation.218 In response, creditors
sued.
Four cases challenging the constitutionality of the joint resolution
reached the Supreme Court in January 1935. Two of the cases involved
private railroad obligations; the other two involved U.S. government debt.
The press closely followed their path; they were front-page material in the
New York Times and the Wall Street Journal.219 They were the subject of
numerous law review articles leading up to the argument,220 and were
foremost among the president’s preoccupations. Roosevelt described the
cases in terms of essential sovereign prerogatives.221 He briefly considered
ways of pressuring the Court to uphold the joint resolution and prepared a
scathing radio address to deliver in the event of an adverse ruling.222
217. Text of the Two Reports on the Gold Resolution, N.Y. TIMES, May 30, 1933, at 2.
218. Joint Resolution to Assure Uniform Value to the Coins and Currencies of the United States,
H.R.J. Res. 192, 73d Cong. (1933) (known as the “Gold Clause resolution”).
219. See, e.g., Second Gold Fight in Supreme Court: RFC Appeals Case to Test Validity of
President’s Decree Voiding Clause in Contracts; Bond Payment at Stake, N.Y. TIMES, Nov. 6, 1934, at
2.
220. See John P. Dawson, The Gold Clause Decisions, 33 MICH. L. REV. 647, 676 n.57 (1935)
(citing ten articles published in the run-up to the Supreme Court argument, all predicting that the joint
resolution would be sustained).
221. ARTHUR M. SCHLESINGER, JR., THE POLITICS OF UPHEAVAL, 1935–1936, at 256 (First
Mariner Books 2003) (1960). Administration advocates referred to the litigants as people who wanted
“$1.69 for their dollar.” Id. Characteristically, the creditors painted a different picture: a $1000 bank
deposit paid back at an arbitrary discount. Norman C. Norman, Letters to the Editor, Our Gold
Certificates, N.Y. TIMES, June 1, 1933.
222. SCHLESINGER, supra note 221, at 257–58; KATHLEEN M. SULLIVAN & GERALD GUNTHER,
CONSTITUTIONAL LAW 24 (15th ed. 2004) (providing an excerpt from the drafted radio address).
Roosevelt’s worries were misplaced. Chief Justice Hughes delivered

the Court’s opinion on February 18, 1935, ruling unequivocally in favor of
the government on private contracts, and only nominally against it in the
cases involving redenomination of federal government bonds.223 The
private contract opinion was far reaching. Hughes rejected out of hand the
creditors’ arguments on retroactive regulation, takings, and due process
grounds.224 He specifically refused to characterize the gold measure as a
constitutional emergency, which might have fixed its duration.225
Instead, the Chief Justice framed the government’s power to rewrite
private contracts as incidental to the entire macroeconomic policy remit,
relying on earlier decisions that interpreted broadly the power to coin and
regulate the value of U.S. currency.226 The central argument in the opinion
reinforced what was then, and continues to be, a widely held view227: that
private contracts must not be read to interfere with legitimate public
Roosevelt’s Fireside Chat of March 9, 1937, in which he proposed his “Court-packing” plan, began by
recounting the narrow 5-4 majority by which the gold clause legislation was upheld, and argued that it
was too dangerous for reforms to depend on a single vote. See FDR’S FIRESIDE CHATS 85 (Russell D.
Buhite & David W. Levy eds., 1992).
223. See Perry v. United States, 294 U.S. 330, 358 (1935); Norman v. Balt. & Ohio R.R., 294 U.S.
240 (1935). See also Nortz v. United States, 294 U.S. 317 (1935); United States v. Bankers Trust Co.,
294 U.S. 240 (1935). We focus on private contracts in this Article. The decisions on government debt
said that the United States did in fact repudiate its debt, but the plaintiffs suffered no damage because
they had no use for gold and, thanks to deflation, had lost no purchasing power. See Perry, 294 U.S. at
358; Nortz, 294 U.S. at 329–30. The opinion’s peculiar reasoning attracted the bulk of the commentary
in the aftermath of the Gold Clause Cases. For a prominent contemporary criticism of the government
debt decisions, see generally Hart, supra note 211. For a more recent view, see Dam, supra note 200, at
518–25.
224. See generally Norman, 294 U.S. 240 (rejecting the plaintiffs’ retroactivity, takings, and due
process claims).
225. See id. at 308–09 (grounding the decision on Congress’s power to regulate commerce
without addressing the parties’ arguments on the status of emergency measures).
226. See, e.g., id. at 303 (“The broad and comprehensive national authority over the subjects of
revenue, finance and currency is derived from the aggregate of the powers granted to the Congress,
embracing the powers to lay and collect taxes, to borrow money, to regulate commerce with foreign
nations and among the several States, to coin money, regulate the value thereof, and of foreign
coin . . . .”).
227. See, e.g., Dawson, supra note 220, at 676 n.57 (citing the general expectation that the Court
would uphold the resolution); John Dickinson, The Gold Decisions, 83 U. PA. L. REV. 715, 715 (1935)
(“The decision in the private bond cases was distinctly conservative . . . .”). See also Richard D.
Friedman, Switching Time and Other Thought Experiments: The Hughes Court and Constitutional
Transformation, 142 U. PA. L. REV. 1891, 1924 (1994) (“For the Justices that had constituted the
majority in Blaisdell, [Norman] was an easy case.”); Seth P. Waxman, The Physics of Persuasion:
Arguing the New Deal, 88 GEO. L.J. 2399, 2416 (2000) (arguing that the private contracts cases were
the easiest Gold Clause Cases to argue). For an example of a modern-day view of the federal
government’s power to regulate private contracts, see Pension Benefit Guaranty Corp. v. R.A. Gray &
Co., 467 U.S. 717, 732–34 (1984).
policymaking.
There is no constitutional ground for denying to the Congress the power
expressly to prohibit and invalidate contracts although previously made,
and valid when made, when they interfere with the carrying out of the
policy it is free to adopt. . . . To subordinate the exercise of the Federal
authority to the continuing operation of previous contracts would be to
place to this extent the regulation of interstate commerce in the hands of
private individuals and to withdraw from the control of the Congress so
much of the field as they might choose by “prophetic discernment” to
bring within the range of their agreements.228
Although the decision was widely expected, the other side offered its
share of drama. Justice McReynolds, “his face set and red, his high-pitched
southern voice quivering with cold anger,”229 spoke for the four
dissenters.230 In extemporaneous remarks later reported in the Wall Street
Journal, he declared the episode a “confiscation of private rights
and . . . repudiation of national obligations,” compared Roosevelt to Nero
for both his monetary indiscipline and despotism, and pronounced the
Constitution “gone.”231
But Wall Street seemed to see things differently. As soon as the
Court’s decision was announced, both stocks and bonds rallied—perhaps
recognizing the need to remove a debt overhang, or perhaps relieved to see
the end of at least some of the uncertainty of the previous year.232
Over time, the Gold Clause Cases on private contracts came to
reinforce the proposition that Congress had broad power to rewrite such
contracts where they interfered with otherwise lawful federal policies.233
This power was particularly expansive with respect to the macroeconomy.
228. Norman, 294 U.S. at 309–10. Cummings was yet more colorful:
[T]hese gold contracts have invaded the federal field. It is not a case of federal activity
reaching out into a private area. So obsessed are our opponents by the idea of the sanctity of
contracts that they are even prepared to assert their validity when they preëmpt the federal
field. To me this seems a monstrous doctrine. These claimants are upon federal territory. They
are squatters in the public domain, and when the Government needs the territory they must
move on.
Cummings Argument, supra note 199, at 257.
229. SCHLESINGER, supra note 221, at 260.
230. See Norman, 294 U.S. at 316 (McReynolds, J., dissenting).
231. LEUCHTENBURG, supra note 197, at 144; SCHLESINGER, supra note 221, at 260; Justice
McReynolds’ Remarks on Gold Case Decision, WALL ST. J., Feb. 23, 1935, at 1.
232. See Kroszner, supra note 209, at 10–25 (tracing the progression of the Gold Clause Cases
and evaluating the impact of the gold clause abrogation).
233. See, e.g., Seese v. Bethlehem Steel Co. Shipbuilding Div., 168 F.2d 58, 62 (4th Cir. 1948)
(citing Norman to refute a claim that the government’s interference with a private contract constituted a
taking).
It was used to render perfectly sensible contract terms unenforceable where

their ubiquity alone was perceived to harm the public.234 The substance of
potentially offending terms was secondary—specifying it in advance would
require “prophetic discernment” (a bon mot from an earlier case quoted by
all sides and the court in the gold episode).235
Yet for purposes of overcoming contract rigidity, the lesson of the
Gold Clause episode is narrow. The joint resolution and the jurisprudence it
inspired targeted a single boilerplate clause that had populated a large
portion of long-term debt contracts in the United States. A similar measure
might help address formal rigidity in securitization contracts where it is a
function of a similarly distinct and widespread term (for example, an
outright prohibition on underlying mortgage modification). It would do
little to overcome structural or functional rigidity.
B. UTILITY HOLDING COMPANIES AND STRUCTURAL RIGIDITY
Our next example, the Public Utility Holding Company Act of 1935
(“PUHCA”), addresses structural rigidity. PUHCA was an antitrust and
industrial policy initiative: a New Deal strike against “bigness” with a dose
of consumer protection.236 Title I of the law sought to break up a system of
holding companies that had controlled most of the power transmission in
the country using complex and highly leveraged corporate structures.237
PUHCA required utility holding companies to register with the SEC and, in
the so-called death sentence provision, gave the SEC authority to break up
registered companies whose structures it deemed uneconomic or
unnecessarily complex.238 Although PUHCA is known primarily as an
instance of successful deconcentration, our interest lies in the law’s
effectiveness at countering financial complexity, leverage, and regulatory
arbitrage.
234. See Norman, 294 U.S. at 313–14; Second Gold Fight, supra note 219.
235. See, e.g., Norman, 294 U.S. at 310.
236. LEUCHTENBURG, supra note 197, at 156–57 (quoting Louis Brandeis); SCHLESINGER, supra
note 221, at 307 (quoting Sen. Burton K. Wheeler’s remarks introducing the Wheeler-Rayburn bill that
became PUHCA). See also JOEL SELIGMAN, THE TRANSFORMATION OF WALL STREET 247 (1995)
(referring to PUHCA’s enforcement as “the most effective antitrust enforcement program in United
States history”).
237. See SCHLESINGER, supra note 221, at 302–05; SELIGMAN, supra note 236, at 127–31.
238. See Public Utility Holding Company Act (PUHCA) of 1935, Pub. L. No. 74-333, § 11(b), 49
Stat. 803, 820–21 (repealed 2005); SCHLESINGER, supra note 221, at 306. Early drafts of the law
compelled dissolution; the version that passed put the burden on the SEC to prove that the structures
were “unduly or unnecessarily complicate[d]” and served no useful purpose. PUHCA § 11(b), 49 Stat.
at 820–21; SCHLESINGER, supra note 221, at 306, 308.
By 1932, thirteen corporate groups controlled three-quarters of private

electricity generation in the United States; the three largest held over 40
percent.239 The groups were structured as pyramids to maximize control
and leverage. At the top were financial and service firms, which held small
controlling blocks of stock in further tiers of holding companies, which in
turn held controlling positions in operating companies. Each company in
the pyramid was capitalized overwhelmingly with bonds and nonvoting
stock, which gave a small group of industrialists and financiers with
common stock at the top of the pyramid effective control of the entire
system in exchange for what was often a tiny investment.240 Classified
boards, interlocking directorates, and voting trusts completed the
governance picture. Congressional committee and independent commission
reports also revealed massive accounting irregularities and scandalous
intragroup loan and transfer pricing schemes.241
In an age of stark rhetoric, the utilities attracted some of the starkest.
“Evil” and “holding companies” stuck together in official and private
statements, from Roosevelt to Will Rogers.242 Judge Robert E. Healy,
counsel to the Federal Trade Commission, described the holding company
system as “a parasite and excrescence”; then–Tennessee Valley Authority
director David Lilienthal called it a “tapeworm”; and the term “private
socialism” appeared often in the debates of the day to condemn the
antidemocratic power concentration in the utilities’ hands.243 Such
sentiments were motivated by blatant fraud, operational inefficiencies, and
effective exploitation of the rate-paying populace by skimming from the
top of the pyramid. Other factors driving the outrage in 1935 evoke today’s
securitization debates more directly: financial leverage, complexity, and the
concomitant escape from regulation.
239. SCHLESINGER, supra note 221, at 303–04; SELIGMAN, supra note 236, at 127. Schlesinger
writes of control; Seligman describes sixteen holding companies with “ownership interests” in over 90
percent of the nation’s private electrical output. The statistics in both accounts come from three reports
on the utility holding companies. The Federal Trade Commission, the House Interstate Commerce
Committee, and the National Power Policy Committee all investigated the holding company problem in
great detail. Their reports, published in 1935, formed the basis for the legislation. SELIGMAN, supra note
236, at 128–30.
240. SELIGMAN, supra note 236, at 127–28 (citing FED. TRADE COMM’N, UTILITY
CORPORATIONS, Part 72-A, at 136–54 (1935)). At the extreme, an investment of $23,000 could buy
control of a $1.2 billion empire. Id. at 128.
241. SCHLESINGER, supra note 221, at 303–04; SELIGMAN, supra note 236, at 128–29.
242. See SCHLESINGER, supra note 221, at 305, 312. Schlesinger recounts Roosevelt’s Freudian
slip in the 1935 State of the Union address: where his text read “abolition of the evil features of holding
companies,” the president said “abolition of the evil of holding companies.” Id. at 305.
243. See id. at 303–05, 307, 310, 312.
Excessive leverage was blamed for a disproportionately high

bankruptcy rate among utilities following the stock market crash of
1929.244 The utilities had argued that debt financing was essential for
investment, and that the operating companies could not have secured
enough capital to supply power to the country on their own, without
geographically diversified funding vehicles.245 In this view, a diversified
enterprise had the capacity to carry more debt and would be more resilient
in a downturn. But by the late 1920s, investment slowed246 while the failure
rate flew in the face of the diversification argument. W. M. W. Splawn,
special counsel to the congressional committee investigating the holding
companies, criticized their financing methods as borderline sham: they
“manufacture nothing . . . except securities.”247 Several years after
PUHCA’s enactment, in a speech geared to promote its enforcement,
William O. Douglas (then chairman of the SEC) called “[d]iversity of
risk . . . merely a slogan for the security salesman.”248
Convoluted financial organization made the leverage problem
intractable. General Electric chairman Owen D. Young admitted privately
that the utility structures were “so complicated that I feel sure that most of
the men responsible for operating them were misled by their own
mechanisms.”249 Wendell Willkie, then president of the Commonwealth &
Southern utility group, sought to distinguish financial engineering—the
“technical development” of the holding company device250—from the few
bad apples who abused it during “a crazy period [the 1920s] when men
went crazy and did a lot of foolish things.”251 He failed.
244. SELIGMAN, supra note 236, at 127–28. The Supreme Court later framed the control, leverage,
and financial speculation problems as linked:
Most of the financing of the various companies in the structure occurred through the sale to
the public of bonds and preferred stock having low fixed returns and generally carrying no
voice in the managements. . . . [A] relatively small but strategic investment in common
stock . . . in the higher levels of a pyramided structure often resulted in absolute control of
underlying operating companies . . . . A tremendous “leverage” in relation to that stock was
thus produced; the earnings of the top holding company were greatly magnified by
comparatively small changes in the earnings of the operating companies. The common stock
of the top holding company might quickly rise in value and just as quickly fall, making it a
natural object for speculation and gambling.
Am. Power & Light Co. v. SEC, 329 U.S. 90, 101–02 (1946) (footnote omitted).
245. See SCHLESINGER, supra note 221, at 308; SELIGMAN, supra note 236, at 179–82.
248. William O. Douglas, Address to the American Bar Association (July 26, 1938), excerpted in
SELIGMAN, supra note 236, at 181.
250. Id. at 320.
251. Id. at 308. Schlesinger characterized Willkie’s effort as an attempt to present the holding
company as “an affable, inarticulate giant.” Id. at 320.
Financial, legal, and organizational complexity, achieved “through

every tactic which could occur to the fertile minds of Sullivan &
Cromwell,” made the holding companies effectively impervious to
regulation.252 Simplification was a central and plainly stated object of
Title I of PUHCA, which dealt with the holding companies.253 The Act
sought to eliminate “undue and unnecessary” structural complexity because
it was bad as such, because it begot inequity, and because it took economic
and financial activity outside the state’s purview. The Supreme Court later
amplified this sentiment, interpreting PUHCA’s death sentence section “to
prevent the use of [interstate commerce] channels to propagate and
disseminate the evils which had been found to flow from unduly
complicated systems and from inequitable distributions of voting
power.”254
Once Roosevelt and his circle decided to get rid of the holding
companies, they chose between two models. Treasury Department officials
proposed taxing the companies out of existence, imposing an intercompany
dividend levy.255 The National Power Policy Committee proposed
administrative dismantling.256 PUHCA embodied the second model.
Although it gave the SEC discretion to simplify, it also required the
elimination of “great-grandfather” structures—where three or more tiers of
holding company subsidiaries separated the top and the bottom of the
corporate pyramid.257
It is useful to compare PUHCA with the later approach of the
Investment Company Act of 1940,258 which similarly targeted financial
complexity and abuse. The Investment Company Act was a deliberately
moderate measure, designed to pass a hostile Congress.259 As such, it
limited itself to prospective regulation. PUHCA was emphatically
retrospective, much like the Gold Clause resolution. It dismantled existing
contracts and corporate structures that had been perfectly legal but now
252. Id. at 304.

253. Public Utility Holding Company Act (PUHCA) of 1935, Pub. L. No. 74-333, § 1(c), 49 Stat.
803, 804 (repealed 2005).
254. Am. Power & Light Co. v. SEC, 329 U.S. 90, 100 (1946).
255. Cf. Michael S. Barr & James A. Feldman, Ctr. for Am. Progress, Issue Brief: Overcoming
Legal Barriers to the Bulk Sale of At-Risk Mortgages (Apr. 2008), available at
http://www.americanprogress.org/issues/2008/04/pdf/reimc_brief.pdf (proposing the use of tax
penalties to incentivize mortgage modifications).
257. See PUHCA § 11(b)(2), 49 Stat. at 821.
258. Investment Company Act of 1940, 15 U.S.C. §§ 80a-1 to 80a-52 (2006).
259. See SELIGMAN, supra note 236, at 222, 226–27.
stood in the way of federal policy.

PUHCA’s tumultuous birth foreshadowed its equally dramatic
enforcement path, which spanned two decades.260 Opponents of the law
prophesied financial chaos and economic meltdown, mass liquidation,
national paralysis, and the financial ruin of widows and orphans.261 The
utilities first resisted registration, repeatedly, and often successfully,
challenging PUHCA’s constitutionality in the lower courts, where
Republican appointees still held sway.262 It took intricate maneuvering for
the SEC to get the Supreme Court to uphold the constitutionality of the
registration provisions only.263 The risk of testing the Section 11(b) death
sentence was too high in the late 1930s; it had to wait until more judges
sympathetic to the New Deal came onto the bench.264
In the meantime, successive SEC chairmen implemented PUHCA
flexibly to avoid a court challenge.265 Official outreach and cajoling failed
miserably at first. Before enforcement began in earnest around 1940 (as the
courts were becoming safe territory for the SEC), the utilities generally
failed to take advantage of Section 11(e) of PUHCA, which allowed them
to design reorganization plans of their own.266 But by the early 1940s, the
credible threat of judicial enforcement combined with the SEC’s adaptable
administrative enforcement strategy helped break the logjam. In particular,
the SEC secured the cooperation of common stock holders—the most
junior claimants who stood to lose everything in reorganization—with tax
breaks and distribution formulas based on going-concern values. This
allowed them to recover even where preferred stock did not get paid in full.
Analysts reported that both common and preferred stock rose after filing
Section 11 plans.267
260. See id. at 127–38, 181–83, 218–22, 241–64.

261. See SCHLESINGER, supra note 221, at 308, 310.
262. See SELIGMAN, supra note 236, at 131–38, 250.
263. See Elec. Bond & Share Co. v. SEC, 303 U.S. 419 (1938).
264. See SELIGMAN, supra note 236, at 250.
265. See id. at 179–80.
266. Public Utility Holding Company Act (PUHCA) of 1935, Pub. L. No. 74-333, § 11(e), 49
Stat. 803, 822 (repealed 2005) (providing that a holding company or a subsidiary of a holding company
could “submit a plan . . . for the divestment of control, securities, or other assets, or for other
action . . . for the purpose of enabling such company or any subsidiary company thereof to comply with
the provisions” required under the death sentence section).
267. See SELIGMAN, supra note 236, at 252–55, 257. Otis & Co. v. SEC, 323 U.S. 624 (1945),
rejected a challenge to the SEC’s policy of ignoring absolute priority in Section 11 cases. The Supreme
Court held that the term “liquidation” both in the contract and in bankruptcy law did not apply to
Section 11 proceedings. It also held—like the Norman Court before it—that contracts “cannot be
permitted to operate” in the aftermath of contrary federal legislation. See id. at 638.
It bears emphasis, however, that the catalytic effect of SEC actions

was obtained in the context of significantly stepped-up enforcement and
more wide-ranging court challenges. Two cases brought by the subsidiary
holding companies of Electric Bond & Share Co. affirmed the
constitutionality of the death sentence in 1946. In North American Co. v.
SEC, the Court ruled that ownership and distribution of holding company
securities across state lines subjected the companies to federal regulation
under the Commerce Clause; that PUHCA constituted a reasonable
regulation, reasonably implemented; and that an SEC divestment order was
not a taking under the Fifth Amendment.268 American Power & Light Co.
v. SEC was decided several months later. It rejected a vagueness challenge
to Section 11 standards; in particular, it held that the standards prohibiting
“unduly or unnecessarily complicate[d]” structures were not vague. The
Court also reemphasized that Congress was “completely uninhibited” by
the Commerce Clause in choosing its regulatory means to override
previously lawful contracts.269 Echoing North American, it observed that
“the federal . . . power is as broad as the economic needs of the nation.”270
The brief opinion used the word “evil” at least a dozen times.
But as with the Gold Clause Cases, hostilities subsided as time went
on, and the business community began to see the benefits of PUHCA.
According to contemporary observers, “There is increased recognition that
these steps in the enforcement of the Act have been akin to a surgical
operation, through which the dead skin (the top holding company) was
being cut away from the pores (the operating companies) in order to allow
the latter to breathe.”271 The surgery—PUHCA’s administrative
mandates—succeeded at countering not only anticompetitive industrial
concentration, but also financial complexity, leverage, and regulatory
arbitrage.
Complexity, leverage, and regulatory arbitrage are often cited among
the risks to financial stability from asset securitization. PUHCA’s approach
to overcoming these threats also holds lessons for overcoming structural
rigidity in modern contracts even as it reveals the extent of the challenge.
Where the obstacles to contract modification lie in complex legal
structures, reorganizing may take years and considerable technical and
268. N. Am. Co. v. SEC, 327 U.S. 686 (1946).

269. Am. Power & Light Co. v. SEC, 329 U.S. 90, 100 (1946).
270. Id. at 104.
271. Robert M. Blair-Smith & Leonard Helfenstein, A Death Sentence or a New Lease on Life? A
Survey of Corporate Adjustments Under the Public Utility Holding Company Act, 94 U. PA. L. REV.
148, 201 (1945) (internal quotation marks omitted).
political savvy. It is much harder than cutting isolated terms out of

boilerplate contracts. Sustained administrative mandates with ample
discretion here replaced the surgical legislative strike of the Gold Clause
resolution.
C. FARM MORTGAGES: A FAILED RESPONSE TO FUNCTIONAL RIGIDITY
The lessons of PUHCA and the Gold Clause episode for overcoming
contract rigidities are essentially optimistic. Not so with the New Deal
experience in fighting the farm mortgage crisis. That experience illustrates
the intractable challenge of functional rigidity in the form of creditor
coordination problems. New Deal policymakers attempted to deal with the
crisis through bankruptcy law and encouraging voluntary workouts, but
they failed because they never managed to address the power of secured
creditors.
Between 1929 and 1933, one-third of American farmers lost their
farms in foreclosure; 200,000 Americans lost farms in 1933 alone.272 The
farm foreclosure crisis of the Great Depression triggered massive federal
intervention in agriculture, agriculture finance, and rural development.273
Among the key pieces of New Deal farm legislation was the Emergency
Farm Mortgage Act of 1933,274 which authorized the Federal Land Banks
(“FLBs”)275 to refinance farm mortgage loans on generous terms,276 and
272. See Lee J. Alston, Farm Foreclosures in the United States During the Interwar Period, 43 J.
ECON. HIST. 885, 886–88 & tbl.1 (1983). See also MICHAEL L. COOPER, DUST TO EAT: DROUGHT AND
DEPRESSION IN THE 1930S, at 28 (2004); DAVID E. HAMILTON, FROM NEW DAY TO NEW DEAL:
AMERICAN FARM POLICY FROM HOOVER TO ROOSEVELT, 1928–1933, at 148–69 (1991) (describing the
farm credit crisis); VAN L. PERKINS, CRISIS IN AGRICULTURE: THE AGRICULTURAL ADJUSTMENT
ADMINISTRATION AND THE NEW DEAL, 1933, at 10–35 (1969) (same); THEODORE SALOUTOS & JOHN
D. HICKS, AGRICULTURAL DISCONTENT IN THE MIDDLE WEST: 1900–1939 (1951).
273. See, e.g., Civilian Conservation Corps Act, Pub. L. No. 75-163, 50 Stat. 319 (1937) (creating
the Civilian Conservation Corps); Rural Electrification Act of 1936, Pub. L. No. 74-604, 49 Stat. 1363
(creating the Rural Electrification Administration); Soil Conservation and Domestic Allotment Act,
Pub. L. No. 74-46, 49 Stat. 163 (1935); Federal Farm Mortgage Corporation Act of 1934, Pub. L. No.
73-88, 48 Stat. 344; Authority Act of 1933, Pub. L. No. 73-17, 48 Stat. 58, upheld by Ashwander v.
TVA, 297 U.S. 288, 339–40 (1936); Emergency Farm Mortgage Act of 1933, Pub. L. No. 73-10, 48
Stat. 41 (repealed 1947); Agricultural Adjustment Act of 1933, Pub. L. No. 73-10, 48 Stat. 31.
274. Emergency Farm Mortgage Act, 48 Stat. 41.
275. The Federal Farm Loan Act of 1916, Pub. L. No. 64-158, 39 Stat. 360 (repealed 1971),
established twelve FLBs in districts that mirrored those of the Federal Reserve System. The FLBs were
supervised by the Federal Farm Loan Board, which was comprised of the Treasury secretary and four
other members. Id. § 3, 39 Stat. at 360–62. One of these members served as the Board’s executive and
the Farm Loan Commissioner. Id. The FLBs’ stock could be held by state and federal governments and
private entities, but a substantial portion was to be held by hundreds of National Farm Loan
Associations (“NFLAs”), which were also created by the Federal Farm Loan Act. See id. § 7, 39 Stat. at
365–67. The NFLAs’ membership and stock ownership were restricted to farm mortgage borrowers. Id.
temporarily guaranteed the interest on the FLBs’ bonds to ensure

refinancing liquidity.277 The refinancing, however, had to be a first
mortgage and could not be for more than the amount of the unpaid
principal, up to 50 percent of the value of the land plus 20 percent of the
buildings thereon.278 Farmers could also get additional low-cost loans from
the Land Bank Commissioner, secured by a first or second mortgage on
their realty and personalty.279 Farmers were limited, however, to a
maximum 75 percent loan-to-value refinancing on farm land, and FLBs’
and Land Bank Commissioner’s loans were only available if all senior
mortgage debt was retired. In other words, the refinancing had to satisfy all
§ 8, 39 Stat. at 367–68. An NFLA could be formed by any group of ten or more natural persons owning
or about to own farm land that could serve as collateral for an FLB mortgage. Id. § 7, 39 Stat. at 365–
67.
In the FLB system, farmers obtained loans from their local NFLA (or if there was none, from an
FLB agent or even directly). To obtain an FLB loan via the NFLA, a farmer had to purchase stock equal
to 5 percent of the loan in one of the NFLAs. Id. § 8, 39 Stat. at 367–68. The stock served as collateral
for the loan, in addition to the mortgaged property. Id. § 11, 39 Stat. at 369–70. NFLAs were permitted
to take only first mortgages in farm property. The NFLA funded the loan by borrowing from the local
FLB. To do so, the NFLA purchased stock in the local FLB equal to 5 percent of the loan. Id. § 7, 39
Stat. at 365–67. This stock served as collateral, in addition to the mortgage on the farm property. When
the loan was repaid, the NFLA’s stock in the FLB was retired, as was the borrower’s stock in the
NFLA. Id. §§ 7–8, 39 Stat. at 365–68.
The FLBs funded their loans to the NFLAs by issuing tax-exempt, mortgage-backed bonds
through the Federal Farm Loan Board. See id. § 26, 39 Stat. at 380 (exemption from taxation). An FLB
that wished to raise capital outside of a stock sale would provide mortgages as security to the Federal
Farm Loan Board, which would issue bonds against the mortgages. Id. § 18, 39 Stat. at 375–76. The
FLB providing the mortgage collateral for the bonds would be primarily liable on the bonds, but there
was a cross-guaranty by the other FLBs. Id. § 21, 39 Stat. at 377–78. There was no federal guaranty of
the bonds whatsoever, although the federal government regulated the bonds’ issuance. As the major
assets of the FLBs were farm mortgages, these bonds were essentially farm mortgage-backed securities.
In addition to the FLBs, the Federal Farm Loan Act created joint-stock land banks. These were
federally chartered entities that had virtually the same powers as FLBs, but in which no government
investment was permitted. Id. § 16, 39 Stat. at 374–75. The joint-stock land banks were limited to
operating in two contiguous states. Id. The joint-stock land banks issued bonds via the Federal Farm
Loan Board just like FLBs, but the bonds were physically distinguished. Id.
276. ARCHIBALD M. WOODRUFF, JR., FARM MORTGAGE LOANS OF LIFE INSURANCE COMPANIES
138–39 (1937). The refinanced loans were to be at reduced interest rates and with amortization
payments deferred for five years. Id. at 139.
277. Id. at 139. The FLBs were permitted to issue bonds with interest backed by the U.S.
government until the sooner of the yield on their bonds being no more than 4 percent or two years after
issuance. Id.
278. Id.
279. Id. These loans were at 5 percent interest, with amortization commencing three years after
issuance. Emergency Farm Mortgage Act § 32, 48 Stat. at 48. The amortization period was between ten
and twenty years. See id. The loans were originally for a maximum of $5000. Id. The lending limit was
raised in 1934 to $7500. Federal Farm Mortgage Corporation Act of 1934, Pub. L. No. 73-88, sec. 10,
§ 32, 48 Stat. 344, 347. The Farm Loan Commissioner’s loans were financed by a loan from the
Reconstruction Finance Corporation. Emergency Farm Mortgage Act § 32, 48 Stat. at 48.
prior liens. As the leading monograph on the topic observes:

In virtually all cases, the sum of the Land Bank and the Commissioner’s
loans more than covered the amount of the first mortgage. In badly
distressed cases, however, there frequently were several mortgages, first,
second, and third, in addition to tax delinquencies, miscellaneous
judgments, and other debts. In such cases the reduction of principal was
essential to Federal refinancing. The question was who was to take the
loss.280
Although there was no problem of formal contractual rigidity for
1930s farm mortgages, there was a collective action problem among
assorted farm creditors, who could not agree who should incur the write-
down loss necessary for the farmer to refinance. The response to this
collective action problem was to rely on state law creditors’ compositions
and extensions: voluntary state law arrangements under which a debtor’s
creditors collectively agreed to a structure of debt forgiveness. But under
state law, no creditor could be compelled to accept less than what it was
owed. Collective action problems ensued; state law composition attempts
often failed. Extraordinary state resources were brought in to no avail:
To facilitate the adjudication of debt revision, the Farm Credit
Administration requested the governors of the states to appoint special
committees to bring borrowers and creditors together. . . . Since the
committees lacked authority to compel reductions, many cases went
through hearing after hearing and much wrangling before any
adjudication could be effected.281
Federal bankruptcy law provided a possible mechanism to overcome
coordination problems: imposing composition terms on nonconsenting
creditors. Thus, in 1933, Congress enacted Section 75 of the Bankruptcy
Act, entitled “Agricultural Compositions and Extensions.”282 Under this
280. WOODRUFF, supra note 276, at 145. This book was published as the winner of Williams
College’s David A. Wells, Class of 1847 Prize for best economics essay, subject to the curious
restriction that
[n]o subject shall be selected for competitive writing on investigation and no essay shall be
considered which in any way advocates or defends the spoliation of property under form or
process of law; or the restriction of Commerce in times of peace by Legislation, except for
moral or sanitary purposes; or the enactment of usury laws; or the impairment of contracts by
the debasement of coin; or the issue and use by Government of irredeemable notes or
promises to pay intended to be used as currency and as a substitute for money; or
which defends the endowment of such “paper,” “notes” and “promises to pay” with the legal
tender quality.
Walter Buckingham Smith, Foreword to WOODRUFF, supra note 276, at iii. This Article would not
qualify.
281. WOODRUFF, supra note 276, at 145–46.
282. Act of 1933, Pub. L. No. 72-420, § 75, 36 Stat. 1467, 1470–73 (codified at 11 U.S.C. § 75
(1933)). The law had a sunset provision of 1938. Id. § 75(c), 36 Stat. at 1473.
provision, “conciliation commissioners” were created in rural areas as

adjuncts of federal courts. Financially distressed farmers could file
petitions for debt adjustment.283 The adjustment could be a composition,
which entailed fractional payment of the unsecured debts, or an extension,
which entailed the postponement of the debt’s maturity date.
To achieve an agricultural composition or extension, a majority vote
of creditors, in both dollar amount and number, was required,284 and in the
case of a composition, priority claims (including secured debt, tax claims,
and wages) had to be paid in full. An agricultural composition or extension
could, however, extend the maturity date of any type of debt.285 The
composition or extension was to be confirmed if the court found that the
composition or extension was an “equitable and feasible method of
liquidation for secured creditors and of financial rehabilitation for the
farmer.” The court also had to find that it was in the “best interests of all
creditors,” proposed and accepted in good faith, and not usurious.286
The compositions failed to deal with farmers’ financial distress. From
June 1934 to June 1937, there were only about 1700 compositions
approved out of nearly 13,700 petitions filed.287 Whether the other petitions
resulted in voluntary arrangements is unclear. While little is known about
the operation of agricultural compositions and extensions under Section 75,
the dual majority vote requirements (dollar amount and number) gave
mortgage creditors an effective veto over any composition, as they likely
held the majority of debt by dollar amount. The veto held by secured
creditors meant that restructuring efforts were inevitably hobbled,
especially when the problematic debt was the secured mortgage.
In stark contrast to the SEC’s flexible treatment of priorities in
PUHCA liquidations, the strict implementation of Section 75 to satisfy
secured creditors rendered the section largely incapable of restructuring
farm mortgages. It could help farmers with their other debts, but this was of
little use for a farmer whose primary financial obligation was a mortgage.
Recognizing the shortcomings of Section 75, Congress passed the
Frazier-Lemke Farm Bankruptcy Act of 1934.288 The Act sought to create a
283. See id. § 75(c), 36 Stat. at 1473.

284. Id. § 75(g), 36 Stat. at 1472.
285. Id. § 75(j), 36 Stat. at 1472.
286. Id. § 75(i), 36 Stat. at 1472.
287. ALBERT GAILORD HART, DEBT AND RECOVERY: A STUDY OF CHANGES IN THE INTERNAL
DEBT STRUCTURE FROM 1929 TO 1937 AND A PROGRAM FOR THE FUTURE 143 (1938).
288. Frazier-Lemke Farm Bankruptcy Act of 1934, Pub. L. No. 73-486, 48 Stat. 1289, invalidated
by Louisville Joint Stock Land Bank v. Radford, 295 U.S. 555 (1935) (current version at 11 U.S.C.
legal mechanism for farm reorganization that would cover secured debt and
prod voluntary compositions. The Frazier-Lemke Act of 1934 permitted a
farmer who could not obtain the requisite consents for a composition to
declare bankruptcy, at which point the farmer’s property—excluding
property exempt at state law—passed into the bankruptcy estate subject to
any liens.289 The Act then gave the farmer two options.
First, if the affected secured creditors consented, the farmer could
purchase property back from the estate at its newly appraised value by
making deferred payments over six years at 1 percent interest with a
balloon payment for 85 percent of the value at the end of the contract.290
The appraisal was to be at “fair and reasonable value, not necessarily the
market value.”291 The farmer gained title and possession to the property
immediately,292 and the completion of the payments discharged all claims,
including mortgage debt.293 Thus, the farmer could repurchase his estate for
its depressed market value, even if it was less than the value of the debt.294
Alternatively, if the affected secured creditors did not consent, the
bankruptcy court was required to stay all proceedings for five years, during
which time the debtor retained possession of the property if it paid a
reasonable annual rental fee.295 At the end of the five years, the debtor
could redeem the property by paying the newly appraised price of the
property.296
Thus, the choice for a secured creditor was between five years of
rental payments based on a reduced property valuation, after which the
debtor had the option to purchase, or six years of backloaded purchase
price payments, also based on the newly reduced valuation. There was little
functional difference between these choices. Both were essentially a right
for the debtor to rent for five years at reduced rates. If the debtor could
afford market rental rates, the debtor could retain the property at his or her
§ 203(a) (1940)).
289. Id.
290. Id. § 3, 48 Stat. at 1290.
291. Id. § 1, 48 Stat. at 1289.
292. Id. § 2, 48 Stat. at 1289–90.
293. Id. § 6, 48 Stat. at 1290.
294. This is essentially what is allowed in a redemption under § 722 of the Bankruptcy Code. 11
U.S.C. § 722 (2006). In redemption, the debtor need only pay the allowed secured claim, which cannot
be greater than the fair market value of the property. See id. §§ 506(a)(1), 722. Redemption, however,
does not allow for deferred payments, which makes it of little value to most debtors, as they are not
flush with cash. The Frazier-Lemke Act, in contrast, restored the income-producing farm to the debtor
immediately and allowed the debtor to use the farm’s income to redeem the farm.
295. Frazier-Lemke Act of 1934 § 7, 48 Stat. at 1291.
296. Id.
will.297
Less than a year after the enactment of the Frazier-Lemke Act, the
Supreme Court struck down the law as an unconstitutional taking.298 A
scaled-down version of the Act, the Frazier-Lemke Act of 1935, was
passed three months later.299 The revised version, upheld by the Supreme
Court,300 functioned mainly as an extension of the equity of redemption
allowed by many states’ foreclosure laws.301
So long as the mortgagor’s equity of redemption had not been
foreclosed, the 1935 Frazier-Lemke Act permitted it to be extended for up
to three years, during which time the debtor retained possession of the
property, but the Act required the debtor to make rental payments.302 At the
end of the three years, the debtor could redeem the property for its
appraised “reasonable market value.”303 A secured creditor could, however,
insist on a public auction of its collateral, at which it could credit bid,304
subject to the debtor having ninety days to redeem the property at the
auction sale price plus 5 percent annual interest.305 If a mortgagee credit
bid at the sale, the redemption price would be the amount of the mortgage
debt outstanding.
At best, then, the revised Frazier-Lemke Act gave a farm debtor three
years of living as a renter with an option to buy, but only by paying off the
full mortgage. Compared with the original 1934 Frazier-Lemke Act, the
1935 version gave the debtor less time to buy back the farm for more
money. Creditors suffered a delay, but avoided a write-down.
Nevertheless, the promise of federal subsidy, combined with the
existence of foreclosure moratoria in many states306 and the successive
iterations of the Frazier-Lemke Act, spurred voluntary compositions in
which creditors took write-downs to the 75 percent “cumulative-loan-to-
297. Cf. U.S. Sav. Ass’n of Tex. v. Timbers of Inwood Forest, 484 U.S. 365 (1988) (finding that
an undersecured creditor was not entitled to postpetition interest—economically equivalent to market
rents—on foreclosure value of property).
298. Louisville Joint Stock Land Bank v. Radford, 295 U.S. 555 (1935).
299. Frazier-Lemke Act of 1935, Pub. L. No. 74-384, 49 Stat. 942.
300. Wright v. Vinton Branch of Mountain Trust Bank of Roanoke, 300 U.S. 440 (1937).
301. E.g., Home Bldg. & Loan Ass’n v. Blaisdell, 290 U.S. 398, 447–48 (1934); Roe, supra note
132.
302. Frazier-Lemke Act of 1935 §§ 5–6, 49 Stat. at 943–45.
303. Id. §§ 4, 6, 49 Stat. at 942–45.
304. Louisville Joint Stock Land Bank v. Radford, 295 U.S. 555, 594–95 (1935).
305. Frazier-Lemke Act of 1935 § 6, 49 Stat. at 943–45.
306. HART, supra note 287, at 140–41 (noting that twenty-two states passed moratorium laws
preventing foreclosure sales).
value” ratio necessary to permit government-supported refinancing. A

major post-Depression study of debt adjustment noted:
A creditor who blocked a settlement by refusing to accept a scale-down
lost his chance to exchange his claim for [federally guaranteed] Federal
Farm Mortgage Corporation bonds. Eventually the creditor might
recover by a private settlement or by foreclosure; but in view of
moratorium laws, depressed values of farm lands, and the expensiveness
of long-drawn-out proceedings[,] many creditors preferred the certainty
of part payment to the possibility of ultimate full collection.307
Yet despite the appearance of progress, there was no significant
decline in the average rate of farm foreclosures until after 1940.308 Given
that state foreclosure moratoria already offered the delay imposed by the
1935 Frazier-Lemke Act, it appears that few farmers saw much advantage
in filing for bankruptcy. Relatively few farm bankruptcy petitions were
filed between 1934 and 1937. Around 10,500 farmers filed for bankruptcy,
and there were 2,350,000 mortgaged farms in 1935,309 a year when 30
percent of the FLBs’ farm mortgages were in default or foreclosure.310
Both versions of the Frazier-Lemke Act essentially provided the farmer-
debtor with more time to recover financially, but neither version reduced
debt burdens. The redemption provisions in the Act were of little use to
farmers who lacked cash, and the delay provided did little to add to that
from states’ foreclosure moratoria.311 Bankruptcy, in the limited form
available in the 1930s, did not provide the solution to the farm mortgage
crisis of the Depression.
In retrospect, it looks like the New Deal approach to farm foreclosures
lacked both the simplicity of the Gold Clause resolution and PUHCA’s
combination of draconian law and administrative discretion. Unlike the
Gold Clause resolution, the farm foreclosure initiatives operated on a more
complex series of contracts and had to juggle a number of competing
interests. Unlike PUHCA, the farm foreclosure initiatives did not strive for
grand industrial restructuring for the long term, but sought to stem
immediate losses. No farm initiative could have afforded to delay
implementation until it could face a friendlier Supreme Court. In the end,
the New Dealers settled on a fairly conventional sticks-and-carrots
approach to address intractable collective action problems in the context of
307. Id. at 154.

308. HAMILTON, supra note 272; PERKINS, supra note 272; SALOUTOS & HICKS, supra note 272.
309. HART, supra note 287, at 143.
310. Id. at 139 fig.12.
311. See generally Home Bldg. & Loan Ass’n v. Blaisdell, 290 U.S. 398 (1934).
plummeting farm prices. By and large, they failed.
D. FARM MORTGAGES REVISITED: CHAPTER 12 OF THE BANKRUPTCY

CODE
The limitations of New Deal farm relief were a lesson learned, and as
a result, the federal response to the next major farm finance crisis in the
1980s specifically targeted collective action problems involving multiple
secured creditors.
The 1970s were a boom time for American agriculture. Farmers
borrowed heavily to expand their operations to meet the demand for U.S.
agricultural exports, particularly to the Soviet Union.312 Inflation helped, as
it raised the prices of farm products and boosted the value of farmland
while keeping farm debts constant. The anti-inflationary policies of the
1980s and declining demand for U.S. agricultural exports left many farmers
overleveraged, having expanded on credit to meet demand that no longer
existed. By 1985, between two and three hundred thousand farmers were
on the brink of failure.
Farmers frequently had multiple secured creditors, who often failed to
agree on restructuring terms.313 Fully secured creditors were more likely to
push for liquidation, while undersecured creditors tended to favor
restructuring.
Congress responded in 1986 by restructuring the farm credit system
yet again. In addition, however, Congress enacted Chapter 12 of the
Bankruptcy Code, which permitted family farmers and fishermen to modify
their mortgages.314 Prior to 1986, the Bankruptcy Code was not well suited
to small farmers who wished to keep their farms. Chapter 7 requires
surrendering the farm,and many farmers are ineligible for Chapter 13
repayment plans because of their large secured farm debts.315 Furthermore,
Chapter 11 reorganization was too complicated and expensive, and it gave
creditors a veto over the reorganization plan. Chapter 12 copied many of
the features of Chapter 13, but removed the eligibility requirements that
prevented farmers from using it, and restricted its application to family
farmers and fishermen. Chapter 12 was initially enacted with a sunset
312. Farm Credit Administration, History of FCA and the FCS, http://www.fca.gov/about/history/
historyFCA_FCS.html (last visited Sept. 1, 2009).
313. See NEIL E. HARL, THE FARM DEBT CRISIS OF THE 1980S, at 275 (1990).
314. See 11 U.S.C. § 1222(b)(2) (2006).
315. See 11 U.S.C. § 109(e) (creating secured and unsecured debt limits as part of Chapter 13
eligibility).
provision but was made permanent in 2004.316 While Chapter 12 could not
restructure American agriculture, it was successful at creating a mechanism
through which farmers could renegotiate secured debt and avoid
preventable foreclosures.
The experience of the 1930s exposed ways in which secured credit
created functional rigidities in farm mortgage contracts. Functional
obstacles to modification are grounded in the economics of financial
transactions; they do not always have a clear legal predicate in contract or
statute. A comprehensive bankruptcy regime designed to reflect the
economics of farm finance—rather than to target any particular contract
term or legal structure—was an imperfect fix, but it offers the most
promising one to date for a foreclosure epidemic rooted in financial
complexity and contract rigidity.
VII. CONCLUSION: BREAKING THE SPELL
We began this Article with a survey of contractual rigidities embedded

in the design of RMBS PSAs. These contracts are unusual because they
layer formal restraints on modification, legal structures that serve as
obstacles to renegotiation, and economic incentives that make renegotiation
functionally difficult or impossible. Market participants appear to perceive
these features as barriers to modification and have cited them as reasons for
not restructuring. Formal, structural, and functional rigidities together can
produce a near-immutable contract317—a rare creature ostensibly barred by
contract and bankruptcy doctrine, one that combines the features of
mundane commercial promises and complex financial products in an effort
to make itself impervious to individual credit risk.
We have also outlined ways of dealing with these three categories of
rigidity. Formal rigidity presents the simplest problem. The Gold Clause
episode and the Supreme Court jurisprudence it produced suggest that it
would be relatively simple to legislate away both the contractual and TIA
barriers to amending RMBS PSAs. Narrowly targeted legislation could
make the clauses unenforceable on public policy grounds. Indeed, a
proposal along these lines was mooted in Congress in early 2009.318
316. Originally, Chapter 12 had a sunset of September 30, 1993. It was later extended to a sunset
of September 30, 1998, and then made permanent.
317. See Davis, supra note 128.
318. See TARP Reform and Accountability Act of 2009, H.R. 384, 111th Cong. § 205
(“Notwithstanding any other provision of law, and notwithstanding any investment contract between a
servicer and a securitization vehicle or investor, a servicer—(i) shall not be limited in the ability to
modify mortgages, the number of mortgages that can be modified, the frequency of loan modifications,
The Gold Clause episode also highlights the fact that rigidity as such
is not a problem in our analysis; rather, it is pervasive rigidity that can harm
communities, drive down asset prices, and depress growth. Thus, instead of
banning or rewriting antimodification provisions, the state might prefer to
sell a limited number of “rigidity licenses”319 to give parties that value
them the benefits revealed in the theoretical literature,320 while preventing
the harmful spillover effects from pervasive rigidity.321 Licensing comes
closest to offering ex ante predictability, which may be particularly useful
in designing financial arrangements to succeed old-style securitization. But
banning, rewriting, and licensing share a limitation: all three are hard to
take beyond formal rigidity. Where the obstacle to modification is not a
readily identifiable, ubiquitous term, but a varied, potentially numerous,
and occasionally amorphous set of structures and practices, a surgical strike
may be impractical: too hard to specify ex ante and too broad or too narrow
ex post. A varied and amorphous practice is effectively unquantifiable,
hence difficult to control through licensing.
Contractual and statutory triggers are another way of addressing the
effects of ubiquitous rigidity in advance. Contracting parties that seek
rigidity to obtain information and to discipline counterparties get neither of
these benefits when the entire economy is in a free fall: good agents and
debtors look just like the bad ones, and no amount of discipline can bring
back vanished markets and cashflows. Therefore, parties should have no
trouble agreeing that their antimodification devices do not apply beyond a
specified set of market or macroeconomic parameters. A purely contractual
mechanism of this sort would evoke elements of indexation discussed
earlier, and of a traditional force majeure clause. The state can also
prescribe or imply such triggers in all or some categories of contracts,
which would be a form of countercyclical regulation of contracts.322 In
or the range of permissible modifications . . . .”). Cf. Mayer et al., supra note 194, at 417 (proposing
legislative mortgage contract modification).
319. We are grateful to Ian Ayres for this insight.
320. See, e.g., Jolls, supra note 128, at 203–06, 236–37.
321. Note that the reasons for banning antimodification provisions are not limited to our
arguments about spillover effects. See, e.g., Beatty v. Guggenheim Exploration Co., 122 N.E. 378, 381
(N.Y. 1919) (invalidating an antimodification clause essentially on freedom of contract grounds—
which can cut both ways).
322. Cf. Marcus Miller & Joseph Stiglitz, Bankruptcy Protection Against Macroeconomic Shocks:
The Case for a “Super Chapter 11” (Dec. 1999) (unpublished manuscript), available at
http://www2.warwick.ac.uk/fac/soc/csgr/research/keytopic/global/milrstig.pdf (recommending across-
the-board “Super Chapter 11” debt relief triggered by exchange rate devaluation); Posner & Zingales,
supra note 2 (recommending ex post mortgage modification in line with property value reductions by
zip code). See generally RICHARD E. SPEIDEL, CONTRACTS IN CRISES (2007) (discussing the application
of the contractual doctrine of excuse in response to government interference with contracts in crisis).
either case, the nub of the problem becomes identifying a trigger that is
sufficiently precise and that properly balances public and private concerns.
A trigger that is too generous would eviscerate the antimodification device
even in good times and make it worthless for the parties; a trigger that is
too stingy would not solve the rigidity problem and its spillover effects.
The PUHCA experience reveals the utility of combining muscular
legislation with administrative discretion in countering structural rigidity.
In the RMBS context, legislation could impose recovery values on
securitization vehicles, mandate simplification of financial structures, and
provide for the regulatory and accounting treatment of the result. An
administrative agency—possibly, but far from obviously, the SEC—might
be charged with determining specific restructuring terms and negotiating
with the industry representatives. Here too, congressional authority is well
established to impose financial reorganization to advance reasonable public
policy goals, such as financial stability and revival of the housing market.
We are left with functional rigidity. The experience with restructuring
secured farm debt in the 1930s suggests that this is the most formidable
kind of contract rigidity, susceptible at best to comprehensive treatment in
bankruptcy. Enforcement suspensions and public subsidies may have
helped on the margins, but they did not go a long way to solve what proved
to be intractable coordination problems.
Perversely, the fact that RMBS PSAs combine all three layers of
rigidity may be cause for optimism. We do not believe that either
legislative amendment on the Gold Clause resolution model, or
administrative mandates of the PUHCA sort, would solve today’s mortgage
securitization dilemma outright—although both might help to some extent.
Bankruptcy might offer a more durable solution, but it looks like the most
politically daunting of the three options since it is, by definition, permanent
and comprehensive rather than temporary and narrow.323 However, while
bankruptcy is comprehensive in its coverage of obligations and creditors of
any given debtor, as far as the debtors are concerned, it is still a case-by-
case device. This creates a problem where failure is ubiquitous because of
an economic downturn: the judicial and administrative systems may not
have the capacity to process all the bankruptcy cases in time to mitigate the
323. On April 30, 2009, the Senate voted 45-51 against an amendment to S. 896 (111th Cong.),
the Helping Families Save Their Homes Act of 2009, which would have allowed principal residence
mortgage modification in Chapter 13 bankruptcy. The House version of the bill, H.R. 1106, had a
bankruptcy modification provision, but due to the amendment’s defeat in the Senate, the final law did
not provide for bankruptcy modification. See Helping Families Save Their Homes Act of 2009, Pub. L.
No. 111-22, 123 Stat. 1632. The final version was passed on May 20, 2009.
spillover effects of rigidity. Here too, contractual and statutory

modification triggers described earlier may serve as safety valves before or
outside bankruptcy.
Apart from any particular solution, unless public policy addresses the
rigidities head on, the U.S. economy and financial system will remain
inextricably linked to a species of Frankenstein contracts designed to be
impervious to subsidy and suasion. Reducing rigidity may not work alone,
but it may alter the incentives calculus for securitization participants and
facilitate other measures to address the foreclosure crisis. Moreover, the
very threat of government intervention to rewrite contracts that the parties
claim are too rigid to modify may act as an incentive to find private routes
to modification. Contracts that claimed to be immutable in 2008 might
acquire surprisingly flexible interpretations a year later, in the legislative
shadow.
Even so, overcoming rigidity in the manner of the New Deal
initiatives we describe is costly: it may undermine confidence in contract
enforcement, and if used frequently or incautiously, it may undermine faith
in the rule of law. Yet while contract enforcement is a central tenet of the
free market system, the law has long countenanced exceptions in the form
of bankruptcy, antitrust law, consumer protection, and many other forms of
economic regulation. When a contract becomes a suicide pact, the Supreme
Court has said repeatedly that Congress may rewrite it.
It is important to leave the decision to rewrite contracts in the hands of
the legislature; the executive may find it too easy and tempting to use. We
can only hope that abrogation is a tool that will be used sparingly and
responsibly after considered debate. If the legislature abuses it, everyone
will pay the price.324 But if it intervenes in a targeted way in response to
real cataclysms, we expect that faith in the law will survive. It might, in
fact, grow stronger when combined with faith in the government’s ability
to protect society from Frankenstein contracts. Financial stability depends
not only on parties’ ability to bind themselves with enforceable contracts,
but also on the government’s ability to regulate—and sometimes to
rewrite—contracts that threaten society.
324. See Horacio Spector, Constitutional Transplants and the Mutation Effect, 83 CHI.-KENT L.
REV. 129, 144–45 (2008) (arguing that Argentina’s experience rewriting contracts every ten years is
unlike the United States’ experience rewriting them every hundred).

FRANKENSTEIN MORTGAGES-PSAs-CONTRACTS - Securitization & Mortgages!!

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

FRANKENSTEIN MORTGAGES-PSAs-CONTRACTS - Securitization & Mortgages!!

Uploaded by

Copyright:

Available Formats

DO NOT DELETE 12/2/2009 3:34 AM

Modification-proof contracts boost commitment and can help

* Associate Professor of Law, American University Washington College of Law.

1076 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 82:1075

financial markets, and the macroeconomy.

2009] REWRITING FRANKENSTEIN CONTRACTS 1077

B. RIGID CONTRACTS IN BANKRUPTCY THEORY ........................ 1118

I. INTRODUCTION: CONTRACT ALCHEMY

1. MARY WOLLSTONECRAFT SHELLEY, FRANKENSTEIN; OR, THE MODERN PROMETHEUS 72

1078 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 82:1075

commitment between parties and helping them overcome information

3. See infra Part IV.A.

2009] REWRITING FRANKENSTEIN CONTRACTS 1079

magical features of financial alchemy (such as bankruptcy remoteness,

1080 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 82:1075

contracts, a comprehensive restructuring mechanism like bankruptcy may

II. THE SPECIES: RMBS

11. Id. at 72.

2009] REWRITING FRANKENSTEIN CONTRACTS 1081

Association (“Fannie Mae”) or the Federal Home Loan Mortgage

1082 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 82:1075

“sponsor”) owns a pool of mortgage loans, which it either made itself

FIGURE 1. Prototypical Private-Label RMBS Transaction

2009] REWRITING FRANKENSTEIN CONTRACTS 1083

and liabilities for bankruptcy remoteness and accounting consolidation

FIGURE 2. Prototypical Private-Label RMBS Transaction: Expanded

23. See infra Part III.C.

1084 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 82:1075

Institutions have many different and overlapping reasons to

2009] REWRITING FRANKENSTEIN CONTRACTS 1085

Figure 3 for an illustration of senior/subordinate tranching. Typically,

FIGURE 3. Tranching of RMBS

1086 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 82:1075

2009] REWRITING FRANKENSTEIN CONTRACTS 1087

III. SECURITIZATION’S RIGIDITIES

Contracts that cannot be modified are more likely to perform

1088 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 82:1075

such a specific return seek to lock in a predictable pattern of behavior on

2009] REWRITING FRANKENSTEIN CONTRACTS 1089

A. FORMAL RIGIDITY 1: PSA LIMITATIONS ON LOAN MODIFICATION

PSA terms often explicitly limit modification of the underlying

1090 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 82:1075

Sometimes modification is forbidden outright;51 sometimes renegotiation is

2009] REWRITING FRANKENSTEIN CONTRACTS 1091

the full cost of modification on the servicer makes them reluctant to

B. FORMAL RIGIDITY 2: STATUTORY AND CONTRACTUAL VOTING

The Trust Indenture Act of 193960 (“TIA”) directly impedes

1092 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 82:1075

requires the consent of 100 percent of their holders. Unanimity is hard to

2009] REWRITING FRANKENSTEIN CONTRACTS 1093

C. STRUCTURAL RIGIDITY: BANKRUPTCY REMOTENESS AND PASSIVE

In corporate bonds, bankruptcy overcomes creditor coordination

1094 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 82:1075

consensually, it can file for bankruptcy and force a debt restructuring.68

68. See id.

2009] REWRITING FRANKENSTEIN CONTRACTS 1095

To make bankruptcy remoteness meaningful, the vehicle must be

1096 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 82:1075

By contracting out of bankruptcy, the securitization SPV creates

81. See supra Part III.B.

2009] REWRITING FRANKENSTEIN CONTRACTS 1097

the originator is often the servicer, overly active management of the

1098 SOUTHERN CALIFORNIA LAW REVIEW [Vol. 82:1075